Secondary Logo

Share this article on:

Clinical Performance and Safety of Closed-Loop Systems: A Systematic Review and Meta-analysis of Randomized Controlled Trials

Brogi, Etrusca MD; Cyr, Shantale PhD; Kazan, Roy MD, MSc; Giunta, Francesco MD; Hemmerling, Thomas M. MSc, MD, DEAA

doi: 10.1213/ANE.0000000000001372
Anesthetic Clinical Pharmacology: Systematic Review Article

Automated systems can improve the stability of controlled variables and reduce the workload in clinical practice without increasing the risks to patients. We conducted this review and meta-analysis to assess the clinical performance of closed-loop systems compared with manual control. Our primary outcome was the accuracy of closed-loop systems in comparison with manual control to maintain a given variable in a desired target range. The occurrence of overshoot and undershoot episodes was the secondary outcome. We retrieved randomized controlled trials on accuracy and safety of closed-loop systems versus manual control. Our primary outcome was the percentage of time during which the system was able to maintain a given variable (eg, bispectral index or oxygen saturation) in a desired range or the proportion of the target measurements that was within the required range. Our secondary outcome was the percentage of time or the number of episodes that the controlled variable was above or below the target range. The standardized mean difference and 95% confidence interval (CI) were calculated for continuous outcomes, whereas the odds ratio and 95% CI were estimated for dichotomous outcomes. Thirty-six trials were included. Compared with manual control, automated systems allowed better maintenance of the controlled variable in the anesthesia drug delivery setting (95% CI, 11.7%–23.1%; percentage of time, P < 0.0001, number of studies: n = 15), in patients with diabetes mellitus (95% CI, 11.5%–30.9%; percentage of time, P = 0.001, n = 8), and in patients mechanically ventilated (95% CI, 1.5%–23.1%; percentage of time, P = 0.03, n = 8). Heterogeneity among the studies was high (>75%). We observed a significant reduction of episodes of overshooting and undershooting when closed-loop systems were used. The use of automated systems can result in better control of a given target within a selected range. There was a decrease of overshooting or undershooting of a given target with closed-loop systems.

Published ahead of print August 1, 2016.

From the *Department of Anesthesia and Intensive Care, University of Pisa, Pisa, Italy; Department of Anesthesia, McGill University, Montreal, Quebec, Canada; and Division of Experimental Surgery, McGill University, Montreal, Quebec, Canada.

Published ahead of print August 1, 2016.

Accepted for publication March 21, 2016.

Funding: Departmental.

The authors declare no conflicts of interest.

Reprints will not be available from the authors.

Address correspondence to Thomas M. Hemmerling, MSc, MD, DEAA, Department of Anesthesia, Montreal General Hospital, C10-153, 1650 Cedar Ave, Montreal, Quebec, Canada H3G1A4. Address e-mail to thomas.hemmerling@mcgill.ca.

A closed-loop system can be defined as an automated control system using the feedback principle. Closed-loop systems are designed to maintain a given variable around a desired set point.

The key elements of these automated control systems are a measurement device (sensor), a controller, and an actuator. The sensor monitors the target parameter and produces a signal (feedback signal) that represents the status of the controlled variable. The controller calculates the difference between the feedback signal and the desired set point and, through different algorithms, produces an output related to the difference between them. Then the actuator converts the digital signal to a physical response (ie, syringe pump delivering the adjusted dose of a given drug).

Since the first introduction of closed-loop anesthesia (in the late 1980s),1,2 several research groups focused on the development and improvement of automated system control.3–8 Closed-loop systems in anesthesia and intensive care seem to achieve an improved control of drug delivery9 and reduce the clinician’s workload.10 In 2014, a Cochrane systematic review and meta-analysis11 on the effectiveness of computerized weaning systems compared with manual control showed a reduction of the duration of weaning, ventilation, and length of stay in the intensive care unit (ICU) in the closed-loop group. Despite all this evidence, automated systems still face great resistance. This phenomenon likely reflects regulatory issues, safety and liability concerns from manufacturers, and the lack of evidence showing a positive clinical impact on patient outcome.

The aim of this systematic review and meta-analysis was to evaluate the clinical performance of closed-loop delivery systems compared with manual control. Clinical performance was defined as the percentage of time that the system was able to maintain a given variable (eg, bispectral index [BIS], SpO2) in a desired target range. The secondary outcome was the evaluation of safety, expressed as the percentage of time or the number of episodes that the system undershoots or overshoots its target value.

Back to Top | Article Outline

METHODS

We performed this review and meta-analysis following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses12 statements. We systematically explored the US National Library of Medicine database (MEDLINE), the Excerpta Medica database (Embase), and Cochrane Central Register of Controlled clinical studies (CENTRAL) for randomized controlled trials (RCTs) from 1999 to 2014. We did not impose any restrictions for language.

The search criteria (in titles, abstracts, and/or all fields) were as follows: (“closed-loop system” or “closed-loop control” or “closed-loop insulin” or “closed-loop anesthesia” or “closed-loop ventilation” or “closed-loop phenylephrine” or “closed-loop oxygenation”) and (“manual” or “automated”). Furthermore, we hand searched the reference lists of the articles to find all relevant articles missed by electronic search. Our search was limited to RCTs on human subjects comparing closed-loop systems with manual control. Reviews, abstracts, conference proceedings, letters to editor, meta-analyses, case reports, and retrospective and preliminary studies were rejected.

We screened all search results, titles, and abstracts of retrieved articles to assess eligibility and then obtained full article copies for all the included studies. Among the publications identified, we excluded studies irrelevant to the topic, nonrandomized studies, technical descriptions, presentation of prototypes without clinical tests, performance test studies, and nonhuman model studies. Trials were also excluded if they compared 2 closed-loop systems and if they did not report our primary outcome. The selected articles were evaluated using the 3-item, 5-point Oxford Quality Scale.13 Studies were rated based on the 3 methodologic features. Two points were given for descriptions of randomization, 2 points for description of the blinding process, and 1 point for the description of withdrawal. We excluded from this review all articles that did not obtain a minimum score of 2.

A data collection form was created with the following main study characteristics: authors, year of publication, number of patients included, type of closed-loop system, controlled variable (ie, BIS, arterial blood pressure [AP], SpO2), and primary and secondary/other outcomes. Two further investigators verified all data. One investigator was in charge of collecting the data and assessing the methodologic validity of all the eligible studies. Our primary outcome was to evaluate the accuracy of automated control versus manual control, expressed as the percentage of time a given variable was maintained in a desired range or the proportion of the target measurements that was within the required range. The evaluation of safety, expressed as percentage of time or the number of the episodes that the controlled variable was above or below the target range, was our secondary outcome.

Data were extracted only from the published articles retrieved and managed in full accordance with the Cochrane Handbook for Systematic Reviews of Interventions 14 and the recommendations of Hozo et al.15. Our primary and secondary outcomes data were then entered into the meta-analysis. When authors stratified the efficacy of maintaining BIS as close to the target into different categories (ie, BIS ≤ 10% excellent; 11%–20% good, 21%–30% poor, and >30% inadequate control), we summed the percentage of excellent and good performance to obtain the total percentage of time that BIS was maintained within the desired range. If the authors investigated the efficacy of 2 algorithms, we included data from both systems. Data expressed as minutes or hours were converted to percentage of time. Data presented as median and ranges were translated to mean and SD following a published formula.15 Publication bias was evaluated by analyzing the funnel plots. We calculated the 95% confidence interval (CI) to summarize continuous data, whereas the odds ratio and 95% CI were calculated for dichotomous data. A random-effects model was applied to analyze the data. Heterogeneity of the retrieved trials was evaluated through the I 2 statistic. I 2 values above 75% reflected a high heterogeneity.16,17 We performed subgroup analyses as sensitivity analysis for the secondary end point. Subgroups were identified in the following manner: the given variable was above the target range; the given variable was below the target range or inadequate control. Meta-analyses and subanalyses were performed when at least 2 trials reported our selected end points. All statistical analyses were performed using the R software (The R Project for Statistical Computing, R version 3.2.1) including a statistical adjustment18 (Hartung-Knapp adjustment for random-effects model19) and Review Manager (RevMan; Computer program. Version 5.3 Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2012).

Back to Top | Article Outline

RESULTS

Electronic literature searches identified 407 articles, and 8 further citations were found by hand searching the reference lists of the studies. We rejected citations on the initial abstract screen for not being pertinent to the topic. This resulted in 44 articles for full text review. Eight further studies were discharged. The reasons for exclusion were as follows: studies comparing 2 different closed-loop systems but not manual control; groups were not randomized; studies did not investigate or report outcomes relevant to this review; and simulation and preliminary studies. Our review finally included 36 randomized clinical trials. Figure 1 illustrates a flowchart of the literature search. Characteristics and outcomes of retrieved articles are listed in Table 1.

Table 1

Table 1

Figure 1

Figure 1

Four studies had a high-quality score of 4 and 1 trial obtained a high-quality score of 5. Twenty-three studies qualified for an intermediate score of 3, and 8 trials had a low-quality score of 2.

Back to Top | Article Outline

Closed-Loop Control for Anesthesia Delivery

Fifteen3,8,10,20,21,27,31,35,38–41,48,50,51 randomized clinical trials comparing closed-loop control for anesthetic drug administration versus manual control were retrieved. Only 1 trial21 involved children. All trials investigated the accuracy of closed-loop systems during the perioperative period with the exception of 1 study50 that evaluated closed-loop systems for sedation in the postoperative period. Another trial35 focused on closed-loop systems for sedation in critically ill patients. The surgical procedures included elective cardiac surgery,20,21,48,50 ambulatory gynecologic procedures,27 elective general surgery,3,8,31,38–40 gynecologic surgery,51 vascular and thoracic surgery,10 and elective spinal surgery.41

A patented closed-loop system called closed-loop anesthesia delivery system for propofol administration was used in 4 trials.3,20,21,50 In another study,48 the same research group tested the clinical feasibility of the updated version of the closed-loop anesthesia delivery system called Improved Anesthetic Agent Delivery System. This new system was able to deliver not only propofol but also volatile anesthetics. Another research group8 evaluated the clinical performance of McSleepy in maintaining a desired level of anesthesia compared with manual control. Differently from the previous systems, this automated system integrates all 3 components of anesthesia (hypnosis, analgesia, and muscle relaxation); it utilizes 3 controlled variables (BIS, analgoscore, and train-of-four ratio) to adjust the drug delivery (propofol, remifentanil, and rocuronium). Dual-loop systems (administrating both propofol and remifentanil) were used in 4 trials.10,35,39,40 In the remaining studies, closed-loop systems for propofol27,31,38,51 or isoflurane41 were utilized.

All studies used BIS as the controlled variable, with the exception of 1 trial40 that utilized M-Entropy. The BIS target range was defined at 50 ± 10 in 11 trials,3,10,20,21,27,35,38,39,41,48,51 at 70 in 1 study,50 and at 45 in 2 trials8,31. The state entropy (SE) target was set in the range of 50 ± 10. All trials presented the percentage of time, and the controlled variables were maintained within the desired target; for 10 trials,3,8,20,21,27,31,35,38,48,50 this outcome was the primary end point.

Figure 2

Figure 2

Compared with manual control, automated systems increased the percentage of time a given variable was maintained in a desired range by 17.4% (95% CI, 11.7%–23.1%; P < 0.0001, Figure 2). Heterogeneity among the studies was significant (I 2 = 0.86). This statistical result indicated that closed-loop systems were able to maintain a given variable within the desired target range for a longer period of time than manual control. In all studies, the percentage of time was higher in the closed-loop group and reached statistical significance in 12 trials.8,10,20,27,31,35,38–40,48,50,51 In 9 trials,8,27,31,35,38–41,51 the authors also evaluated the percentage of time that the controlled variable was above or below (BIS/SE < 40 or BIS/SE > 60) the desired set point. Performing subgroup analyses, we observed more undershooting or overshooting in the manual groups than in the closed-loop groups, and this at −12.3% (95% CI, −18.6% to −6.1%, P = 0.0009). Heterogeneity among the studies was significant (I 2 = 0.96).

Back to Top | Article Outline

Closed-Loop for Insulin Administration

We identified 11 trials22,23,26,29,32,36,42–46 comparing closed-loop insulin delivery system with insulin pump therapy. All automated systems used blood glucose (BG) concentration as the controlled variable.

Three articles23,36,45 evaluated closed-loop insulin delivery systems in the ICU setting. Patients were randomly allocated to receive either an automated insulin infusion protocol or a local care protocol. Standard care consisted of continuous IV insulin infusion,23,45 intermittent subcutaneous insulin administration (CSII),23 or IV insulin bolus regimen.36,45 The BG target range was defined at 100 to 130mg/dL,23 80 to 140 mg/dL,45 and 6 to 8 mM.36 All 3 trials investigated the percentage of time the BG concentration was maintained within the desired target. This was determined as the primary outcome of 2 of these trials.36,45

Compared with manual control, we observed a difference of 24.6% (95% CI, −20.5% to 69.7%; I 2 = 0.96, Figure 3) in the percentage of time a given variable was maintained in the desired range using an automated system. However, this difference was not statistically significant (P = 0.14).

Figure 3

Figure 3

In addition, all articles investigated the number of episodes that the BG was above or below the required target range; 1 trial45 found more hyperglycemic episodes in the standard care group (11 vs 1), whereas no episodes of hypoglycemic events occurred in either group. In another trial,36 the authors did not observe hypoglycemic events in the automated group. One trial23 found a statistically significant reduction of the number of hypoglycemic episodes in the automatic group (P = 0.04). Performing subgroup analyses, we observed that automated systems decreased the percentage of time of inadequate control in comparison to manual control. (Mean difference = -12.3%; 95% CI, -18.6% to -6.1%, P = 0.0009). However, this difference was not statistically significant (P = 0.25)

The remaining 8 articles22,26,29,32,42–44,46 compared the accuracy of closed-loop insulin delivery systems with pump insulin therapy in patients with type 1 diabetes mellitus. Six trials22,26,29,32,42,43 enrolled children. Different types of pumps were used: CSII in 3 trials,22,32,43 sensor-augmented pump therapy in 2 trials,42,44 standard open-loop insulin therapy in 2 trials,26,29 and intraperitoneal open-loop insulin delivery in 1 trial.46 Table 1 illustrates the different target ranges selected for the BG in these trials. Seven23,26,29,32,42–44 of 8 articles were randomized crossover trials; patients were randomly assigned to receive closed-loop insulin infusion or pump therapy and then switched to the other therapy. All trials provided the primary outcome of this review; in 6 trials,22,26,29,32,42,46 this outcome was the primary end point.

Figure 4

Figure 4

Figure 5

Figure 5

Compared with manual control, automated systems increased the percentage of time a given variable was maintained in a desired range by 21.2% (95% CI, 11.5%−30.9%; I 2 = 0.96, P = 0.001 as shown in Figure 4). In all studies, the length of time was longer in the automated control group. This difference reached statistical significance in 6 trials.22,29,32,43,44,46 All articles provided data corresponding to the secondary outcome of this review. In 3 articles,22,42,44 the authors found a statistically significant reduction of episodes of hypoglycemia in the automated group. In the remaining articles, even if not statistically significant, the number of episodes of hypoglycemia or hyperglycemia was higher in the control groups. Performing subgroup analyses, we observed that automated systems decreased the percentage of time of inadequate control (below or above the target range) of BG compared with the manual group by −6.5 % (95% CI, −11.3% to −1.8%, I 2 = 0.92, P = 0.010; Figure 5).

Back to Top | Article Outline

Closed-Loop Control of Ventilation

Figure 6

Figure 6

Eight trials24,25,28,30,33,37,47,52 determined the accuracy of closed-loop systems compared with manual control of ventilation. Newborns were enrolled in 4 trials.24,25,30,52 All studies included patients who required mechanical ventilation or supplemental oxygen. In 6 trials,24,25,28,30,33,52 the study design consisted of 2 consecutive observational periods on automated and manual controls. In the remaining 2 trials,37,47 patients were randomly assigned to either automated or clinician control. One study47 specifically investigated the weaning period. Closed-loop control with an automated FIO2 adjustment was used in 5 trials.24,25,30,33,52 In 3 studies,28,37,47 the authors investigated the feasibility of closed-loop systems for pressure support ventilation to maintain the patient in an acceptable ventilation zone (based on tidal volume, respiratory rate, end-tidal CO2 values). In 5 trials,24,25,30,33,52 the controlled variable was SpO2. In the other 3 trials,28,37,47 respiratory rate, tidal volume, and end-tidal CO2 were used as target variables; desired target ranges are shown in Table 1. In manual control groups, the percentage of time that the given variable remained within the desired range was significantly reduced (mean difference: 12.8%; 95% CI, 1.5%–23.1%; I 2 = 75%, P = 0.03; Figure 6). In all studies, the length of time was longer in the automated control group, and this difference reached statistical significance in 6 trials.24,25,28,30,37,52 Seven trials24,25,28,30,33,37,52 investigated the occurrence of overshoot and undershoot episodes; a statistically significant reduction of episodes of ventilation that was not acceptable28,37 or the percentage of time below30 the target range was found in automated groups in 3 articles; 1 trial33 found that the number of episodes of desaturation was similar between the groups, but the total duration of desaturation was higher in the automated group; 2 trials24,25 found that the number of episodes was almost similar but that the total duration of desaturation was longer in the manual group; in 1 trial,52 the authors observed more episodes of desaturation in the manual group. Performing subgroup analyses, we observed less maintenance of the controlled variable within the target range in the manual group when compared with automated systems by −7.7% (95% CI, −14.2% to −1.2%, P = 0.024). Heterogeneity among the studies was significant (I 2 = 0.97).

Back to Top | Article Outline

Closed-Loop for Vasopressor Administration

Two trials34,49 presented data on the comparison between automated closed loop for vasopressor (phenylephrine or ephedrine) administration and manual control. All articles included patients scheduled for cesarean delivery under spinal anesthesia.

In 1 trial,49 the authors utilized a closed-loop dual-vasopressor automated system. This system administrated phenylephrine to treat hypotension and ephedrine to treat low AP with bradycardia. Beat-to-beat systolic AP (noninvasive AP) was the controlled variable. In the other trial,34 an automated system for phenylephrine administration was used, and AP was the controlled variable. The 2 trials calculated the proportion of systolic measurements that were within ±20% of the baseline.

Compared with manual control, automated systems increased the number of measurements that were within the target range (odds ratio, 1.44; 95% CI, 1.04–2.0; I 2 = 0.77, P < 0.0001). The number of patients with 1 or more episodes of hypotension or hypertension was similar between the manual and automated group in 1 trial.34 In the other trial,49 the number of patients with hypotension was significantly higher in the manual group (P = 0.001).

Back to Top | Article Outline

DISCUSSION

In this systematic review and meta-analysis of randomized clinical trials, we evaluated the accuracy and safety of closed-loop systems compared with manual control. Automated systems increased the length of time that a given variable was maintained in the desired range compared with manual control. This difference reached statistical significance in the majority of trials (28 of 36). This likely suggests that automated systems can obtain a better control of depth of anesthesia, BG level, and ventilation. Because of the limited number of studies gathered, no conclusion could be drawn concerning automated systems for vasopressor administration and for insulin infusion in ICU patients. The incidence of overshooting or undershooting a given control target was reduced during automated control or was at least similar between the groups in all trials reporting this outcome (29 of 36).

Back to Top | Article Outline

Assessment of Clinical Significance of Improvements

To assess the clinical significance of the control aspect of closed-loop systems, we can deduct from this meta-analysis that there is a significant improvement in the percentage of time of “desired” control (parameter in the predefined range of target control). Interestingly, this improvement ranged from 12% for the maintenance of specific ventilation parameters (Figure 6), to 17% for maintenance of a target level of anesthesia (Figure 2) and to 21% of maintenance of a target level of BG level in patients with diabetes mellitus (Figure 4). As an example, for 1-hour duration of anesthesia, the targeted level of anesthesia would be maintained for 10 minutes longer than with manual control: these authors would regard this as a significant clinical advantage.

We chose undershooting or overshooting as a parameter of safety assessment: for example, we would consider undershooting a given BIS target as putting the patients at risk of either waking up or having periods of awareness. The improvement ranged from 7% for avoidance of overshooting or undershooting parameters of ventilation and avoidance of either hypoglycemia or hyperglycemia (Figure 5) to 12% for avoidance of undershooting or overshooting a given level of anesthesia. If one considers that undershooting a given level of anesthesia, too light a level of hypnosis, puts the patient at risk of awareness, whereas overshooting might be related to worse outcome, 5 to 6 minutes less time during 1 hour of anesthesia of exposure to these risks is a significant clinical improvement of closed-loop systems compared with manually administered anesthesia.

We were able to retrieve only 2 meta-analyses on closed-loop systems. Rose et al11 published a Cochrane review and meta-analysis focusing on automated weaning systems. The authors’ first objective was the total duration of weaning from mechanical ventilation with automated weaning systems compared with a noncomputerized strategy in critically ill patients. Further areas of interest were the total duration of ventilation, length of ICU stay, mortality, and the incidence of adverse events between the 2 groups. The data showed a significant reduction in the duration of weaning and total length of mechanical ventilation in automated groups, whereas no statistical differences in ICU stay, mortality, or adverse events were found between groups. In contrast to our study, the authors decided to group the trials retrieved into subcategories based on the automated weaning systems and the weaning strategy used in the control groups (standard care), and then they estimated the CI separately for these subclasses. Regardless of the subgroup analyses, the heterogeneity between the studies was high. The authors highlighted the necessity of high-quality multicenter controlled trials. A Cochrane review and meta-analysis by Burns et al53 investigated the efficacy of closed-loop systems for the discontinuation of mechanical ventilation. This review included 1 RCT with 300 adult postoperative patients and compared automated weaning systems with nonautomated protocols. The data showed a significant reduction in ventilation time to first spontaneous breathing trial, but not in time to successful extubation, total ventilation time, or length of hospital stay. The authors concluded that the evidence was too weak to reach any conclusion on the use of automated weaning systems for the weaning period and that larger high-quality trials were required. The main difference between our review and the aforementioned meta-analyses is that we did not focus our attention on a specific use of closed-loop systems (ie, closed-loop systems for the weaning period) but decided to analyze the overall clinical performance of the automated systems.

There are several limitations to our review. First, we decided to perform a review and meta-analysis focusing on the comparison between automated systems and manual control in different medical fields. As a consequence, we analyzed different kinds of closed-loop systems. We cannot draw any conclusion about the superiority of one technological system compared with another. A second limitation was the limited number of studies retrieved for automated insulin delivery in ICU patients and closed-loop systems for vasopressor administration. Third, the overwhelming majority of trials qualified for an intermediate quality score. Only 1 study49 reported a proper blinding procedure. Another limitation was the high heterogeneity of the studies. This marked heterogeneity can be explained by multiple factors. We included trials that investigated both pediatric and adult patients. Moreover, regarding closed loop for insulin administration, the desired target range for the BG concentration varied widely within the studies (as shown in Table 1). The high heterogeneity likely reflects the different durations of the observational period selected by the authors (eg, 36 hours,29 6 weeks44). The choice of comparing differently designed and engineered automated systems might have increased the heterogeneity.

Another limitation of the review is related to the type of studies published: no studies focused on the influence of closed-loop systems on patient outcome such as reduced length of stay, reduced morbidity, or mortality.

In summary, automated systems increased the length of time that a given variable was maintained in the desired range compared with manual control. The incidence of overshooting or undershooting a given control target was reduced during automated control or was at least similar between the groups in all trials reporting this outcome.

Back to Top | Article Outline

DISCLOSURES

Name: Etrusca Brogi, MD.

Contribution: This author helped conduct the study, analyze the data, and write the manuscript and the archival author.

Name: Shantale Cyr, PhD.

Contribution: This author helped analyze the data and write the manuscript.

Name: Roy Kazan, MD, MSc.

Contribution: This author helped analyze the data and write the manuscript.

Name: Francesco Giunta, MD.

Contribution: This author helped write the manuscript.

Name: Thomas M. Hemmerling, MSc, MD, DEAA.

Contribution: This author helped design the study, conduct the study, analyze the data, and write the manuscript.

This manuscript was handled by: Ken B. Johnson, MD.

Back to Top | Article Outline

REFERENCES

1. Schwilden H, Schüttler J, Stoeckel H. Closed-loop feedback control of methohexital anesthesia by quantitative EEG analysis in humans. Anesthesiology 1987;67:341–7.
2. Schwilden H, Stoeckel H, Schüttler J. Closed-loop feedback control of propofol anaesthesia by quantitative EEG analysis in humans. Br J Anaesth 1989;62:290–6.
3. Puri GD, Kumar B, Aveek J. Closed-loop anaesthesia delivery system (CLADS) using bispectral index: a performance assessment study. Anaesth Intensive Care 2007;35:357–62.
4. Kenny GN, Mantzaridis H. Closed-loop control of propofol anaesthesia. Br J Anaesth 1999;83:223–8.
5. Mortier E, Struys M, De Smet T, Versichelen L, Rolly G. Closed-loop controlled administration of propofol using bispectral analysis. Anaesthesia 1998;53:749–54.
6. Liu N, Chazot T, Trillat B, Pirracchio R, Law-Koune JD, Barvais L, Fischler M. Feasibility of closed-loop titration of propofol guided by the Bispectral Index for general anaesthesia induction: a prospective randomized study. Eur J Anaesthesiol 2006;23:465–9.
7. Simanski O, Janda M, Schubert A, Bajorat J, Hofmockel R, Lampe B. Progress of automatic drug delivery in anaesthesia-the ‘Rostock assistant system for anaesthesia control (RAN)’. Int J Adapt Contr Signal Process 2009;23:504–21.
8. Hemmerling TM, Arbeid E, Wehbe M, Cyr S, Taddei R, Zaouter C. Evaluation of a novel closed-loop total intravenous anaesthesia drug delivery system: a randomized controlled trial. Br J Anaesth 2013;110:1031–9.
9. Hemmerling TM. Automated anesthesia. Curr Opin Anaesthesiol 2009;22:757–63.
10. Dussaussoy C, Peres M, Jaoul V, Liu N, Chazot T, Picquet J, Fischler M, Beydon L. Automated titration of propofol and remifentanil decreases the anesthesiologist’s workload during vascular or thoracic surgery: a randomized prospective study. J Clin Monit Comput 2014;28:35–40.
11. Rose L, Schultz MJ, Cardwell CR, Jouvet P, McAuley DF, Blackwood B. Automated versus non-automated weaning for reducing the duration of mechanical ventilation for critically ill adults and children. Cochrane Database Syst Rev 2014;6:CD009235.
12. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioannidis JPA, Clarke M, Devereaux PJ, Kleijnen J, Moher D. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. BMJ 2009;339:b2700.
13. Moher D, Jadad AR, Nichol G, Penman M, Tugwell P, Walsh S. Assessing the quality of randomized controlled trials: an annotated bibliography of scales and checklists. Control Clin Trials 1995;16:62–73.
14. Higgins JP, Green S. Cochrane Handbook for Systematic Reviews of Interventions. 2008Wiley Online Library.
15. Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and variance from the median, range, and the size of a sample. BMC Med Res Methodol 2005;5:13.
16. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat Med 2002;21:1539–58.
17. Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003;327:557–60.
18. Ledolter J, Dexter F. Analysis of interventions influencing or reducing patient waiting while stratifying by surgical procedure. Anesth Analg 2011;112:950–7.
19. IntHout J, Ioannidis JP, Borm GF. The Hartung-Knapp-Sidik-Jonkman method for random effects meta-analysis is straightforward and considerably outperforms the standard DerSimonian-Laird method. BMC Med Res Methodol 2014;14:25.
20. Agarwal J, Puri GD, Mathew PJ. Comparison of closed loop vs. manual administration of propofol using the Bispectral index in cardiac surgery. Acta Anaesthesiol Scand 2009;53:390–7.
21. Biswas I, Mathew PJ, Singh RS, Puri GD. Evaluation of closed-loop anesthesia delivery for propofol anesthesia in pediatric cardiac surgery. Paediatr Anaesth 2013;23:1145–52.
22. Breton M, Farret A, Bruttomesso D, Anderson S, Magni L, Patek S, Dalla Man C, Place J, Demartini S, Del Favero S, Toffanin C, Hughes-Karvetski C, Dassau E, Zisser H, Doyle FJ 3rd, De Nicolao G, Avogaro A, Cobelli C, Renard E, Kovatchev B; International Artificial Pancreas Study Group. Fully integrated artificial pancreas in type 1 diabetes: modular closed-loop glucose control maintains near normoglycemia. Diabetes 2012;61:2230–7.
23. Cavalcanti AB, Silva E, Pereira AJ, Caldeira-Filho M, Almeida FP, Westphal GA, Beims R, Fernandes CC, Correa TD, Gouvea MR, Eluf-Neto J. A randomized controlled trial comparing a computer-assisted insulin infusion protocol with a strict and a conventional protocol for glucose control in critically ill patients. J Crit Care 2009;24:371–8.
24. Claure N, Gerhardt T, Everett R, Musante G, Herrera C, Bancalari E. Closed-loop controlled inspired oxygen concentration for mechanically ventilated very low birth weight infants with frequent episodes of hypoxemia. Pediatrics 2001;107:1120–4.
25. Claure N, Bancalari E, D’Ugard C, Nelin L, Stein M, Ramanathan R, Hernandez R, Donn SM, Becker M, Bachman T. Multicenter crossover study of automated control of inspired oxygen in ventilated preterm infants. Pediatrics 2011;127:e76–83.
26. Dauber A, Corcia L, Safer J, Agus MS, Einis S, Steil GM. Closed-loop insulin therapy improves glycemic control in children aged <7 years: a randomized controlled trial. Diabetes Care 2013;36:222–7.
27. De Smet T, Struys MM, Neckebroek MM, Van den Hauwe K, Bonte S, Mortier EP. The accuracy and clinical feasibility of a new bayesian-based closed-loop control system for propofol administration using the bispectral index as a controlled variable. Anesth Analg 2008;107:1200–10.
28. Dojat M, Harf A, Touchard D, Lemaire F, Brochard L. Clinical evaluation of a computer-controlled pressure support mode. Am J Respir Crit Care Med 2000;161:1161–6.
29. Elleri D, Allen JM, Kumareswaran K, Leelarathna L, Nodale M, Caldwell K, Cheng P, Kollman C, Haidar A, Murphy HR, Wilinska ME, Acerini CL, Dunger DB, Hovorka R. Closed-loop basal insulin delivery over 36 hours in adolescents with type 1 diabetes: randomized clinical trial. Diabetes Care 2013;36:838–44.
30. Hallenberger A, Poets CF, Horn W, Seyfang A, Urschitz MS; CLAC Study Group. Closed-loop automatic oxygen control (CLAC) in preterm infants: a randomized controlled trial. Pediatrics 2014;133:e379–85.
31. Hemmerling TM, Charabati S, Zaouter C, Minardi C, Mathieu PA. A randomized controlled trial demonstrates that a novel closed-loop propofol system performs better hypnosis control than manual administration. Can J Anaesth 2010;57:725–35.
32. Hovorka R, Allen JM, Elleri D, Chassin LJ, Harris J, Xing D, Kollman C, Hovorka T, Larsen AM, Nodale M, De Palma A, Wilinska ME, Acerini CL, Dunger DB. Manual closed-loop insulin delivery in children and adolescents with type 1 diabetes: a phase 2 randomised crossover trial. Lancet 2010;375:743–51.
33. Johannigman JA, Branson R, Lecroy D, Beck G. Autonomous control of inspired oxygen concentration during mechanical ventilation of the critically injured trauma patient. J Trauma 2009;66:386–92.
34. Kee N, Khaw KS, Ng FF, Tam YH. Randomized comparison of closed-loop feedback computer-controlled with manual-controlled infusion of phenylephrine for maintaining arterial pressure during spinal anaesthesia for caesarean delivery. Br J Anaesth 2013;110:59–65.
35. Le Guen M, Liu N, Bourgeois E, Chazot T, Sessler DI, Rouby JJ, Fischler M. Automated sedation outperforms manual administration of propofol and remifentanil in critically ill patients with deep sedation: a randomized phase II trial. Intensive Care Med 2013;39:454–62.
36. Leelarathna L, English SW, Thabit H, Caldwell K, Allen JM, Kumareswaran K, Wilinska ME, Nodale M, Mangat J, Evans ML, Burnstein R, Hovorka R. Feasibility of fully automated closed-loop glucose control using continuous subcutaneous glucose measurements in critical illness: a randomized controlled trial. Crit Care 2013;17:R159.
37. Lellouche F, Bouchard PA, Simard S, L’Her E, Wysocki M. Evaluation of fully automated ventilation: a randomized controlled study in post-cardiac surgery patients. Intensive Care Med 2013;39:463–71.
38. Liu N, Chazot T, Genty A, Landais A, Restoux A, McGee K, Laloë PA, Trillat B, Barvais L, Fischler M. Titration of propofol for anesthetic induction and maintenance guided by the bispectral index: closed-loop versus manual control: a prospective, randomized, multicenter study. Anesthesiology 2006;104:686–95.
39. Liu N, Chazot T, Hamada S, Landais A, Boichut N, Dussaussoy C, Trillat B, Beydon L, Samain E, Sessler DI, Fischler M. Closed-loop coadministration of propofol and remifentanil guided by bispectral index: a randomized multicenter study. Anesth Analg 2011;112:546–57.
40. Liu N, Le Guen M, Benabbes-Lambert F, Chazot T, Trillat B, Sessler DI, Fischler M. Feasibility of closed-loop titration of propofol and remifentanil guided by the spectral M-Entropy monitor. Anesthesiology 2012;116:286–95.
41. Locher S, Stadler KS, Boehlen T, Bouillon T, Leibundgut D, Schumacher PM, Wymann R, Zbinden AM. A new closed-loop control system for isoflurane using bispectral index outperforms manual control. Anesthesiology 2004;101:591–602.
42. Ly TT, Breton MD, Keith-Hynes P, De Salvo D, Clinton P, Benassi K, Mize B, Chernavvsky D, Place J, Wilson DM, Kovatchev BP, Buckingham BA. Overnight glucose control with an automated, unified safety system in children and adolescents with type 1 diabetes at diabetes camp. Diabetes Care 2014;37:2310–6.
43. Nimri R, Danne T, Kordonouri O, Atlas E, Bratina N, Biester T, Avbelj M, Miller S, Muller I, Phillip M, Battelino T. The “Glucositter” overnight automated closed loop system for type 1 diabetes: a randomized crossover trial. Pediatr Diabetes 2013;14:159–67.
44. Nimri R, Muller I, Atlas E, Miller S, Fogel A, Bratina N, Kordonouri O, Battelino T, Danne T, Phillip M. MD-Logic overnight control for 6 weeks of home use in patients with type 1 diabetes: randomized crossover trial. Diabetes Care 2014;37:3025–32.
45. Plank J, Blaha J, Cordingley J, Wilinska ME, Chassin LJ, Morgan C, Squire S, Haluzik M, Kremen J, Svacina S, Toller W, Plasnik A, Ellmerer M, Hovorka R, Pieber TR. Multicentric, randomized, controlled trial to evaluate blood glucose control by the model predictive control algorithm versus routine glucose management protocols in intensive care unit patients. Diabetes Care 2006;29:271–6.
46. Renard E, Place J, Cantwell M, Chevassus H, Palerm CC. Closed-loop insulin delivery using a subcutaneous glucose sensor and intraperitoneal insulin delivery: feasibility study testing a new model for the artificial pancreas. Diabetes Care 2010;33:121–7.
47. Schädler D, Engel C, Elke G, Pulletz S, Haake N, Frerichs I, Zick G, Scholz J, Weiler N. Automatic control of pressure support for ventilator weaning in surgical intensive care patients. Am J Respir Crit Care Med 2012;185:637–44.
48. Madhavan JS, Puri GD, Mathew PJ. Closed-loop isoflurane administration with bispectral index in open heart surgery: randomized controlled trial with manual control. Acta Anaesthesiol Taiwan 2011;49:130–5.
49. Sng BL, Tan HS, Sia AT. Closed-loop double-vasopressor automated system vs manual bolus vasopressor to treat hypotension during spinal anaesthesia for caesarean section: a randomised controlled trial. Anaesthesia 2014;69:37–45.
50. Solanki A, Puri GD, Mathew PJ. Bispectral index-controlled postoperative sedation in cardiac surgery patients: a comparative trial between closed loop and manual administration of propofol. Eur J Anaesthesiol 2010;27:708–13.
51. Struys MM, De Smet T, Versichelen LF, Van De Velde S, Van den Broecke R, Mortier EP. Comparison of closed-loop controlled administration of propofol using Bispectral Index as the controlled variable versus “standard practice” controlled administration. Anesthesiology 2001;95:6–17.
52. Urschitz MS, Horn W, Seyfang A, Hallenberger A, Herberts T, Miksch S, Popow C, Müller-Hansen I, Poets CF. Automatic control of the inspired oxygen fraction in preterm infants: a randomized crossover trial. Am J Respir Crit Care Med 2004;170:1095–100.
53. Burns KE, Lellouche F, Nisenbaum R, Lessard MR, Friedrich JO. Automated weaning and SBT systems versus non-automated weaning strategies for weaning time in invasively ventilated critically ill adults. Cochrane Database Syst Rev 2014;9:CD008638.
© 2017 International Anesthesia Research Society