PDF icon Download PDF
Endoscopy 2012; 44(06): 584-589
DOI: 10.1055/s-0032-1306776
Original article
© Georg Thieme Verlag KG Stuttgart · New York

Propofol sedation with bispectral index monitoring is useful for endoscopic submucosal dissection: a randomized prospective phase II clinical trial

T. Sasaki
Department of Gastroenterology, Kitasato University School of Medicine, Kanagawa, Japan
,
S. Tanabe
Department of Gastroenterology, Kitasato University School of Medicine, Kanagawa, Japan
,
M. Azuma
Department of Gastroenterology, Kitasato University School of Medicine, Kanagawa, Japan
,
A. Sato
Department of Gastroenterology, Kitasato University School of Medicine, Kanagawa, Japan
,
A. Naruke
Department of Gastroenterology, Kitasato University School of Medicine, Kanagawa, Japan
,
K. Ishido
Department of Gastroenterology, Kitasato University School of Medicine, Kanagawa, Japan
,
C. Katada
Department of Gastroenterology, Kitasato University School of Medicine, Kanagawa, Japan
,
K. Higuchi
Department of Gastroenterology, Kitasato University School of Medicine, Kanagawa, Japan
,
W. Koizumi
Department of Gastroenterology, Kitasato University School of Medicine, Kanagawa, Japan
› Author Affiliations
Further Information

Corresponding author

T. Sasaki, MD PhD
Kitasato University School of Medicine
Department of Gastroenterology/Medical Oncology
Kitasato University East Hospital
2-1-1 Asamizodai Minami, Saganihara
Kanagawa
252-0380 Japan   
Fax: +81-42-7498690   

Publication History

submitted 09 February 2011

accepted after revision 17 January 2012

Publication Date:
25 May 2012 (online)

Also available ateRef
 
 PDF Download Permissions and Reprints

Background and study aims: Endoscopic submucosal dissection (ESD) has become a standard treatment. However, the treatment time tends to be relatively long and insufflation and manipulation of the endoscope can increase pain and discomfort. We aimed to find an optimal method for sedation during ESD.

Patients and methods: Patients scheduled to undergo ESD for early gastric cancer or adenoma were randomly assigned to sedation with midazolam or propofol, and consciousness level was evaluated by bispectral index (BIS) monitoring. Primary end points of effectiveness (three parameters) and secondary end points of safety during ESD and after return to the ward were compared between the groups. Study registration was in the UMIN Clinical Trial Registry (UMIN 000001497), and the institutional trial number was KDOG 0801.

Results: From June 2008 through June 2009, we enrolled 178 patients (90 midazolam, 88 propofol). Regarding safety after ESD, recovery was significantly better in the propofol group immediately after and at 1 hour and 2 hours after return to the ward (P < 0.001). The number of patients who required a continuous supply of oxygen 2 hours after returning to the ward was significantly lower in the propofol group (midazolam 18; propofol 6; P = 0.010). Though propofol seemed to be better for effectiveness and safety, there were no statistically significant differences for all three primary end points and the safety parameters (hypotension, hypoxia, bradycardia).

Conclusions: Propofol with BIS monitoring improved recovery of patients after ESD, though this study was underpowered to prove the effectiveness and safety of propofol.


 

Introduction

The advent of endoscopic submucosal dissection (ESD) has led to remarkable progress in the treatment of early gastric cancer. ESD is a technique for the resection of early gastric cancer that involves three steps: injecting fluid into the submucosa to elevate the lesion, incising the mucosa surrounding the lesion, and dissecting the submucosa beneath the lesion under direct vision. It is performed with newly developed devices, such as an insulation-tipped diathermic knife (IT knife; Olympus Medical Systems, Tokyo, Japan). As compared with conventional endoscopic mucosal resection (EMR), ESD has substantially increased the rate of en bloc resection, even of large lesions that were previously difficult to treat endoscopically [1]. Consequently, the indications for endoscopic treatment have been extended to include early gastric cancers amenable to en bloc resection. Such lesions are considered to have an extremely low risk of lymph node metastasis [2]. The use of ESD has rapidly increased during the past few years, but it is not without drawbacks. ESD requires a high level of technical expertise, as well as a longer operation time [3]. The endoscopic procedure is extremely complex and technically demanding. Moreover, distension, incision, and dissection of the gastric wall can cause intense pain. Higher levels of sedation and analgesia are thus required as compared with conventional endoscopic procedures. Agents such as midazolam have been used for sedation in patients who receive endoscopic therapy. In general, however, the effective dose range of such agents differs considerably among patients, making it difficult to achieve a stable level of sedation. Also the dose is often increased to suppress body movements, leading to oversedation, potentially causing hypoxemia and decreased levels of consciousness after patients return to the hospital ward [4]. Several randomized controlled trials have found that propofol is superior to benzodiazepines when used as a sedative for endoscopic examinations [5] [6]. However, in Japan the safety instructions provided with propofol state that “propofol should be used only by physicians who have extensive experience in anesthetic techniques or intensive care.” This has been interpreted to mean that propofol should only be used by anesthesiologists. Therefore, propofol has rarely been used in patients who receive endoscopic treatment. In particular, whether propofol is useful for sedation in patients who undergo ESD remains to be fully investigated. We conducted this randomized prospective phase II clinical trial (institutional number KDOG 0801) to compare the sedative effectiveness and safety of propofol with those of midazolam in patients with gastric tumors who underwent ESD. During sedation, the level of consciousness was assessed by bispectral index (BIS) monitoring. Our goal was to establish an optimal method for sedation in patients who undergo ESD.


 

Patients and methods

Patients

The inclusion criteria were as follows: (i) early gastric cancer or gastric adenoma for which ESD was indicated; (ii) preoperative histological diagnosis of group III, IV, or V [7]; (iii) age 20 years or older; (iv) American Society of Anesthesiologists (ASA) classification of I or II [8]; (v) adequately maintained organ functions; (vi) no history of allergy to midazolam or propofol; and (vii) written informed consent directly from the patient.

The indications for ESD in patients with early gastric cancer were in accordance with the recommendations of Gotoda et al. [9]: differentiated mucosal cancer unaccompanied by an ulcer (irrespective of tumor size); differentiated mucosal cancer 3 cm or less in diameter accompanied by an ulcer; undifferentiated mucosal cancer 2 cm or less in diameter unaccompanied by an ulcer; or differentiated cancer 3 cm or less in diameter with a submucosal invasion depth of 500 µm or less [9].

Exclusion criteria included: (i) a ASA classification of III, IV, or V; and (ii) history of allergy to midazolam or propofol.


 

Study protocol

The protocol for this single-center randomized prospective phase II clinical trial was first drafted in October 2007. The study was approved by the ethics committee of Kitasato University in June 2008 and was registered with the UMIN Clinical Trial Registry (UMIN 000001497) in Japan before initiation.

After obtaining informed consent, endoscopists registered the participants at a data center. The staff of the data center used the randomization function of Microsoft Excel (RAND) to randomly assign participants to the midazolam group or the propofol group. At the time of randomization, participants were stratified according to age (younger than 70 years vs. 70 years or older) and alcohol intake history (on average one or more occasions per week vs. less than one occasion per week), to take into account the influence of these factors on adverse events or the required magnitude of a sedative dose.

After randomization, the data center staff reported the assignment results to the endoscopists just before the ESD procedure was begun. The ESD procedure was done by experienced operators who had previously performed ESD in at least 30 patients.

The Data and Safety Monitoring Committee was responsible for interim analyses and for stopping trials if needed. An interim analysis by this committee was scheduled for when the first 70 patients had been enrolled in the trial.


 

Sedation

A gastroenterologist with extensive experience in anesthetic techniques or intensive care was responsible for sedation (the physician in charge of anesthetization). The target depth of sedation during ESD was a score of – 3 to – 5 on the Richmond Agitation–Sedation Scale (RASS) and a BIS monitoring index of 60 to 80 (BIS Index Monitor; Aspect Medical Systems Inc., Newton, Massachusetts, USA) [10] [11].

Patients in both groups were given one ampoule of pethidine hydrochloride (35 mg/1 mL) by intravenous injection at the start of treatment. Oxygen was delivered through a nasal catheter at a rate of 2 L/min. Midazolam was diluted with physiological saline solution (1ml = 1 mg) and was slowly injected intravenously at a dose of 0.05 mg/kg to induce sedation. Midazolam was then given in additional doses of 1 to 2 mg each. After the procedure, patients in the midazolam group received flumazenil in a maximum dose of 0.5 mg to promote awakening. In the propofol group, propofol was initially administered at a rate of 2 mL (20 mg)/10 seconds to a maximum dose of 10 mL (100 mg) until sedation was induced; to simulate the current situation in Japan, a target-controlled infusion system was not used. Propofol was then given intermittently in additional doses of 1 to 2 ml (10 to 20 mg) each. In principle, crossover between the sedative groups was not allowed, but patient safety took precedence. Therefore, if the level of sedation remained inadequate after the administration of ≥ 20 mg midazolam (2 ampoules) or ≥ 400 mg propofol (2 ampoules) and the physician in charge of anesthetization judged that the patient’s safety might be compromised, other drugs could be used concurrently.

During ESD, all patients were monitored for electrocardiography, peripheral oxygen saturation, blood pressure (intermittent/tonometric), and BIS index. The physician in charge of anesthetization and nurses ensured the safety of patients. After return to the hospital ward, all patients were monitored for electrocardiography, peripheral oxygen saturation, and blood pressure (intermittent). The patients’ condition was frequently checked by nursing staff.


 

Parameters evaluated

The primary end point of this trial was the effectiveness of propofol sedation with BIS monitoring during ESD, and the secondary end points were patient safety during ESD and in the recovery ward.

Measures of sedative effectiveness were defined as follows: (i) rate of achievement of depth of sedation between – 3 and – 5 (RASS score) throughout the ESD procedure; (ii) lowest RASS score during ESD; and (iii) rate of need for restraint of body movements of patients to avoid accidents and to ensure the safety of the patient during ESD.

The physician in charge of anesthetization recorded the details of other drugs that were used to ensure safety.

Safety during ESD was evaluated on the basis of the presence or absence of hypotension (systolic blood pressure ≤ 90 mmHg), bradycardia (≤ 50 beats per minute), hypoxemia (percutaneous oxygen saturation level ≤ 94 %), arrhythmias, and severe adverse events. Patient safety in the recovery ward was assessed immediately after, 1 hour after, and 2 hours after return to the ward. Hypotension, bradycardia, hypoxemia, the recovery level (RASS score), whether oxygen was administered (i. e., if the percutaneous oxygen saturation level was ≤ 94 %), and adverse events such as falls and aspiration pneumonia were documented.


 

Statistical analysis

We previously expected from our clinical experience with ESD that the rates of achievement of the depth of sedation would be 85 % in the propofol group and 60 % in the midazolam group, and the rates of need for restraint of body movement would be 15 % and 40 %, respectively. To detect each difference between the groups with a significance level of 5 % and a statistical power of 90 % using the chi-square test, we estimated that at least 65 patients would be required in each group. With regard to the lowest RASS score, we estimated that at least 131 patients in two groups combined would be required to detect a 1-point difference between the groups with a significance level of 5 % and a statistical power of 90 % using the Mann–Whitney U test. Taking into account that some patients would be ineligible, we set the target number of patients at 150. However this initial calculation had not taken into account the multiple comparisons among our primary end points; subsequently, when the Bonferroni correction was applied, we considered that for comparing the two groups a significance level of P < 0.0167 was required.

The Mann–Whitney U test was used to compare RASS scores between the groups. The statistical significance of differences between the patient groups regarding demographic characteristics and the incidence of adverse events was assessed using chi-square tests. Differences in procedure times and drug usage were assessed by means of the t test.


 

 

Results

Demographic and clinical characteristics of patients

From June 2008 through June 2009, a total of 178 patients were enrolled; 90 were assigned to the midazolam group and 88 to the propofol group without any exclusions. Median age, sex ratio, and history of alcohol consumption did not differ significantly between the groups. Procedure time, tumor locations, the dose of pethidine hydrochloride, and the maximum oxygen requirement also did not differ significantly between the groups ([Table 1]). To ensure patient safety during ESD, the physician in charge of anesthetization decided to use propofol concurrently in 19 patients, only from the midazolam group (P < 0.001).

Table 1

Sedation for endoscopic submucosal dissection (ESD) of gastric lesions, with bispectral (BIS) index monitoring. Patient, clinical, and procedure details.

Midazolam group
n = 90

Propofol group
n = 88

P value

Age, median (range)

71 (45 – 84)

70 (46 – 83)

0.919

Men/Women, n/n

60 /30

64 /24

0.362

Alcohol history, n

0.654

No

47

43

Yes

43

45

Gastric location

0.478

Upper side

37

44

Lower side

50

42

Both

 3

 2

2

Operation time, mean ± SD, minutes

77.3 ± 37

82.1 ± 36.1

0.387

Drug dose, mean ± SD, mg

15.4 ± 6.8

249.1 ± 128.7

 – 

Drug switch, n

19 patients to propofol, with a mean (SD) propofol dose of 117.4 (96.5) mg

 0

 < 0.001

Pethidine HCI 35-mg ampoules, median (range), n

 1 (1 – 3)

 1 (1 – 3)

0.922

Maximum oxygen dose in each patient, median, L/min

 2 (2 – 8)

 2 (2 – 6)

0.244

 

Sedation effectiveness during ESD

The sedation level during ESD, as assessed on the basis of rate of achievement of depth of sedation between – 3 and – 5 on the RASS score, and on the lowest RASS score, did not differ significantly between the groups. Body movement had to be restricted manually in more patients in the midazolam group (24 patients) than in the propofol group (12 patients, P = 0.030); however, this difference was not significant, because with the Bonferroni correction a significance level of P < 0.0167 was required for the primary end point ([Table 2]).

Table 2

Sedation for endoscopic submucosal dissection (ESD) of gastric lesions, with bispectral (BIS) index monitoring. Sedation levels and safety parameters during ESD.

Midazolam group
n = 90

Propofol group
n = 88

P value

RASS score

Between – 3 to – 5 throughout ESD procedure, n (%)

57 (63.3)

63 (71.2)

0.237[1]

Lowest, median (range)

 – 4 ( – 5 to + 3)

 – 4 ( – 5 to + 2)

0.377[1]

Manual restriction of body movement needed, n (%)

24 (26.7)

12 (13.6)

0.030[1]

Cardiopulmonary parameters, n (%)

Hypotension (systolic blood pressure ≤ 90 mmHg)

31 (34.4)

17 (19.3)

0.037

Bradycardia (pulse rate ≤ 50 /min)

16 (17.8)

21 (23.9)

0.317

Hypoxemia (Spo 2 ≤ 94 %)

16 (17.8)

14 (15.9)

0.739

Severe adverse events

 0 (0)

 0 (0)

– 

RASS, Richmond Agitation–Sedation Scale; Spo 2, percutaneous oxygen saturation.

1 The significance level is P < 0.0167 with the Bonferroni correction.



 

Safety parameters during ESD

During ESD, in the midazolam group, bradycardia and hypoxemia occurred in 16 (17.8 %) and 16 (17.8 %) patients respectively, as compared with 21 (23.9 %) and 14 (15.9 %) patients in the propofol group. These differences were not significant. No patient had any severe adverse event in either group. However, the incidence of hypotension was significantly higher in the midazolam group (31 patients, 34.4 %) than in the propofol group (17 patients, 19.3 %; P = 0.037).


 

Safety parameters during recovery

The proportions of patients who were alert with a RASS score of 0 were significantly higher in the propofol group, immediately after and 1 and 2 hours after return to the ward ([Table 3], [Fig. 1]). The frequencies of hypotension, bradycardia, and hypoxemia after return to the ward did not differ between the groups. However, the proportion of patients who required a continuous supply of oxygen up to 2 hours after return to the ward was significantly higher in the midazolam group ([Table 4]).

Table 3

Sedation for endoscopic submucosal dissection (ESD) of gastric lesions, with bispectral (BIS) index monitoring. Safety parameters after return to the ward: level of consciousness as measured on the Richmond Agitation–Sedation Scale (RASS), and serious adverse events.

Midazolam group
n = 90

Propofol group
n = 88

P value

RASS score after return to the ward

Immediately after

Median (range)

 – 1 ( – 4 to 0)

 0 ( – 5 to 0)

 < 0.001

Score of 0, n (%)

34 (37.8)

62 (70.5)

 < 0.001

1 h after

Median (range)

 – 1 ( – 5 to 0)

 0 ( – 5 to 0)

 < 0.001

Score of 0, n (%)

34 (37.8)

74 (84.1)

0.005

2 h after

Median (range)

 0 ( – 4 to 0)

 0 ( – 1 to 0)

 < 0.001

Score of 0, n (%)

47 (52.2)

85 (96.6)

 < 0.001

Severe adverse events, n (%)

 0 (0)

 0 (0)

– 
Zoom
Fig. 1 Sedation with propofol or midazolam accompanied by bispectral (BIS) index monitoring, during endoscopic submucosal dissection (ESD) of gastric lesions. Recovery immediately after, 1 hour after, and 2 hours after return to the ward, measured on the Richmond Agitation–Sedation Scale (RASS) from – 5 (lowest possible score) to 0.
Table 4

Sedation for endoscopic submucosal dissection (ESD) of gastric lesions, with bispectral (BIS) index monitoring. Patient safety parameters after return to the ward: cardiopulmonary dynamics.

Cardiopulmonary values after return to ward

Midazolam, n (%)
(90 patients)

Propofol, n (%)
(88 patients)

P value

Hypotension (systolic blood pressure ≤ 90 mmHg)

lmmediately

 1 (1.1)

 3 (3.4)

0.306

1 h

 3 (3.3)

 1(1.1)

0.317

2 h

 1 (1.1)

 0 (0)

0.319

Bradycardia (pulse rate ≤ 50 /min)

lmmediately

 6 (6.7)

 8 (9.1)

0.563

1 h

 8 (8.9)

 4 (4.5)

0.244

2 h

 6 (6.7)

 3 (3.4)

0.313

Hypoxemia (Spo 2 ≤ 94 %)

lmmediately

19 (21.1)

13 (14.8)

0.256

1 h

10 (11.1)

 4 (4.5)

0.099

2 h

 5 (5.6)

 4 {4.5)

0.745

Oxygen given 2 h after return

18 (20.0)

 6 (6.8)

0.010

Spo 2, percutaneous oxygen saturation.


 

 

Discussion

ESD is superior to conventional EMR for the treatment of early gastric cancer because it permits a higher rate of en bloc resection and is associated with a lower rate of recurrence. Consequently, ESD is considered a standard treatment in Japan [12]. However, ESD requires greater experience and technical expertise than EMR and has a slightly higher risk of complications such as perforation. The treatment time tends to be prolonged, sometimes exceeding 2 hours. ESD can cause discomfort and pain due to factors such as rotational maneuvers, insufflation, and thermal coagulation extending to the muscularis layer. Among all endoscopic procedures, ESD is thus considered to be the most stressful to both patients and physicians.

ESD requires adequate sedation and pain control. In Japan, a benzodiazepine combined with an opioid is generally used for sedation. At present, ESD is usually performed by a small team of physicians and nurses, not accompanied by an anesthesiologist and without the use of a target-controlled infusion system, BIS monitor, or capnometer.

The use of propofol for ESD has been studied recently [13]. However, currently ESD in Japan is associated with risks, such as an increased incidence of complications caused by procedural mistakes resulting from sudden body movements during endoscopy, increased frequencies of hypotension or hypoxemia due to oversedation, and higher incidences of complications during recovery, including aspiration pneumonia and accidents such as falls.

The present study was designed to simulate the current status of ESD in Japan as closely as possible. We directly compared midazolam and propofol, both which were administered by intermittent intravenous injection, without the use of a target-controlled infusion system. The use of intermittent intravenous injection might be one reason for the lack of significant difference between the groups in the rate of achievement of target sedation. We also used BIS monitoring to ensure patient safety and to prevent oversedation. As a result, there was no significant difference between the groups for safety parameters during ESD except for hypotension. Even though this trial was carried out at only a single center and without blinding, it is nevertheless noteworthy that in the propofol group, more than 70 % of the patients had a RASS score of 0 immediately after returning to the ward. In the midazolam group, however, only 37.8 % of the patients had a RASS score of 0 even at 1 hour after returning to the ward, despite the use of flumazenil in all patients. The most notable difference compared with previously reported studies is the use of BIS monitoring for patient safety, to prevent oversedation during the ESD procedure and to improve recovery of consciousness.

The primary end point of this study was the effectiveness of propofol-induced sedation as assessed by BIS monitoring. However, the study was not double-blinded. A dedicated anesthesiologist managed sedation in all patients on the basis of BIS monitoring, and the type and dose of sedatives could be modified, taking into account patient safety. Because the use of only one variable to evaluate effectiveness might have been inadequate, we decided to concurrently assess three variables. For statistical analysis, the sample size was calculated independently according to each of the three variables. However in our initial calculations we did not take account of the fact that there were multiple comparisons for our primary end points, and subsequently after applying the Bonferroni correction, we considered that a P value of less than 0.0167 was needed to indicate a significant difference between the two groups. As a result, the sample size was too small to show a significant difference in primary end point.

Although the statistical analysis and the sample size were inadequate for demonstrating statistical significance, a P value of 0.030 for comparison between two groups has been considered to be small enough in other clinical trials with large sample sizes. The primary end point and sample size for statistical analysis should thus be reconsidered before proceeding to multicenter clinical trials.

Sedation with propofol has previously been reported to be useful in patients undergoing endoscopic examination [14] [15]. The administration of propofol by specially trained nurses has also been attempted [16] [17]. Among patients who underwent conventional endoscopic examination during sedation with propofol, the proportion who required short-term ventilation with a facemask because of respiratory depression has been estimated to be only 0.1 % [18].

The European Society of Gastrointestinal Endoscopy (ESGE) presented guidelines for nonanesthesiologist administration of propofol (NAAP) in 2010 and stated that, to ensure patient comfort and safety, specific knowledge and skills are necessary for endoscopists and nursing staff using NAAP [19]. Recent guidelines in North America also recommend that propofol should be administered by a specially trained gastroenterologist (gastroenterologist-directed propofol sedation, G-DPS) or a specially trained nurse (nurse-administered propofol sedation, NAPS) for patients who undergo conventional endoscopy of the upper gastrointestinal tract [18]. Propofol-induced sedation has been reported to be very cost-effective when patients undergo procedures such as endoscopic retrograde cholangiopancreatography and endoscopic ultrasonography [20]. In the future, propofol is expected to become an essential sedative for ESD. However, many physicians lack adequate experience in the use of propofol, and the American Society for Gastrointestinal Endoscopy (ASGE) recommends that endoscopists should receive additional training to ensure that propofol is administered safely [21] [22]. Training in sedation and resuscitation techniques such as airway management will become essential for endoscopists in the future.

Several methods can be used to ensure that ESD is performed safely. First, the level of consciousness can be continuously assessed using BIS index monitoring, as in the present study. A BIS monitor is generally available in operating rooms and intensive care units, but is usually not included as standard equipment in endoscopic procedure rooms. In this clinical trial, all patients had BIS monitoring during the procedure, which enabled maintenance of deep sedation during ESD in both the midazolam group and propofol group. The ability to minimize the administered dose of sedative agents by use of BIS monitoring may decrease the incidences of hypotension and hypoxemia caused by oversedation and thereby facilitate early recovery after patients return to the ward. Capnometry has also been reported to be useful for preventing oversedation during ESD [23]. The concurrent use of BIS index assessment and capnometry might allow close monitoring of patients [24].

The types and dosages of concurrently used analgesics remain largely uninvestigated. In this clinical trial, one ampoule (35 mg) of pethidine hydrochloride was given before treatment, and additional doses were given as required. The usefulness of drugs such as fentanyl and ketamine should be studied in the future [25].

ESD is a promising treatment that should be used throughout the world. Although the further development of endoscopists’ skills is important, improved methods for sedation are essential to ensure that ESD can be performed more safely. Unfortunately, it seems that this study was underpowered, but nevertheless our results suggest that propofol is a promising sedative for patients undergoing ESD. Further studies of monitoring techniques and concurrently used analgesics are needed to establish an optimal environment for ESD.


 

 

Competing interests: None.


Corresponding author

T. Sasaki, MD PhD
Kitasato University School of Medicine
Department of Gastroenterology/Medical Oncology
Kitasato University East Hospital
2-1-1 Asamizodai Minami, Saganihara
Kanagawa
252-0380 Japan   
Fax: +81-42-7498690   


   Permissions and Reprints
Zoom
Fig. 1 Sedation with propofol or midazolam accompanied by bispectral (BIS) index monitoring, during endoscopic submucosal dissection (ESD) of gastric lesions. Recovery immediately after, 1 hour after, and 2 hours after return to the ward, measured on the Richmond Agitation–Sedation Scale (RASS) from – 5 (lowest possible score) to 0.