Original article
Capnographic monitoring of midazolam and propofol sedation during ERCP: a randomized controlled study (EndoBreath Study)
Authors
Peter Klare1, Johanna Reiter1, Alexander Meining2, Stefan Wagenpfeil3, Tim Kronshage4, Christoph Geist4, Stefan Heringlake4, Christoph Schlag1, Monther Bajbouj1, Gerhard Schneider5, Roland M. Schmid1, Till Wehrmann6, *, Stefan von Delius1, *, Andrea Riphaus7, *
Institutions
Institutions are listed at end of article.
submitted 11. April 2015 accepted after revision 13. July 2015
Background and study aims: This was to determine whether intervention based on additional capnographic monitoring reduces the incidence of hypoxemia during midazolam and propofol sedation for endoscopic retrograde cholangiopancreatography (ERCP). Methods: Patients (American Society of Anesthesiologists [ASA] I – IV) scheduled for ERCP under midazolam and propofol sedation were randomly assigned to a control arm with standard monitoring or an interventional arm with additional capnographic monitoring. In both arms detection of apnea prompted withholding of propofol administration, stimulation of the patient, insertion of a nasopharyngeal tube, or further measures. The primary study end point was incidence of hypoxemia (oxygen saturation [Sao2] below 90 %); secondary end points included occurrences of severe hypoxemia (Sao2 ≤ 85 %), bradycardia, and hypotension, and sedation quality (patient cooperation and satisfaction).
Bibliography DOI http://dx.doi.org/ 10.1055/s-0034-1393117 Published online: 28.9.2015 Endoscopy 2016; 48: 42–50 © Georg Thieme Verlag KG Stuttgart · New York ISSN 0013-726X Corresponding author: Stefan von Delius, MD Klinikum rechts der Isar der Technischen Universität München II. Medizinische Klinik Ismaninger Str. 22, 81675 Munich Germany Fax: +49-89-41404905 [email protected]
Results: 242 patients were enrolled at three German endoscopy centers. Intention-to-treat analysis revealed no significant reduction in hypoxemia incidence in the capnography arm compared with the standard arm (38.0 % vs. 44.4 %, P = 0.314). Apnea was more frequently detected in the capnography arm (64.5 % vs. 6.0 %, P < 0.001). There were no differences regarding rates of bradycardia and hypotension. Per-protocol analysis showed lower incidence of hypoxemia in the capnography arm compared with the standard arm (31.5 % vs. 44.8 %, P = 0.048). There was one death related to sedation in the standard arm. Sedation quality was similar in the two groups. Conclusion: Intention-to-treat analysis showed hypoxemia incidence was not significantly lower in the additional capnography arm compared with standard monitoring. Additional capnographic monitoring of ventilatory activity resulted in improved detection of apnea. ClinicalTrials.gov identifier: NCT01072474.
Introduction !
Endoscopic retrograde cholangiopancreatography is an invasive procedure that requires the use of sedation [1]. During sedation, cardiopulmonary complications such as apnea and consequent hypoxemia may occur and demand monitoring of the patient. Therefore, electronic monitoring as well as close observation of patient responsiveness and respiratory activity is mandatory [1 – 3]. However, neither clinical observation nor pulse oximetry enable early detection of hypoventilation, apnea, and hypercarbia, or their consequences, including acidosis, catecholamine release, myocardial excitation with arrhythmias, and myocardial depression [2]. On the other
hand, apnea and altered respiration frequently precede hypoxemia with a significant time lag [2, 3]. It has been shown that capnographic monitoring is superior to pulse oximetry and visual assessment for the detection of respiratory depression [4, 5]. Hence, capnography could improve patient monitoring and prevent sedation-related hypoxemia, morbidity, and mortality. The aim of this randomized controlled study was to determine whether early intervention based upon data from capnographic monitoring could reduce the incidence of hypoxemia during midazolam and propofol sedation for ERCP.
* These authors contributed equally to the study.
Klare Peter et al. Capnographic monitoring of midazolam and propofol sedation during ERCP… Endoscopy 2016; 48: 42–50
Downloaded by: University of Pittsburgh. Copyrighted material.
42
Patients and methods !
Study population The study protocol was approved by the local ethics committee and registered at ClinicalTrials.gov (identifier: NCT01072474). Patient enrolment began in February 2010 and concluded in October 2011. Outpatients and inpatients aged > 18 years who were scheduled for ERCP at one of the three participating study sites (II Medizinische Klinik und Poliklinik, Klinikum rechts der Isar der Technischen Universität München, Munich, Germany; Medizinische Universitätsklinik des Knappschaftskrankenhaus der Ruhr-Universität Bochum, Bochum, Germany; and Deutsche Klinik für Diagnostik, Helios Klinik Wiesbaden, Wiesbaden, Germany) were eligible for participation. Exclusion criteria were American Society of Anesthesiologists [ASA] risk classification V, pregnancy, or pre-existing hypoxemia (Sao2 < 90 %), hypotension (systolic blood pressure < 90 mmHg), or bradycardia (heart rate < 50 /min) before the start of the endoscopy. Written informed consent was obtained from all patients.
Study design We conducted a randomized controlled trial at three participating sites in Germany. We assessed the efficacy of early intervention prompted by additional capnographic monitoring (Capnostream 20; Oridion Medical, Needham, Massachusetts, USA) to reduce the incidence of arterial oxygen desaturation (defined as a fall to below 90 %) during ERCP under propofol and midazolam sedation. Patients were randomly assigned to an arm with standard monitoring (“standard arm”) and to an arm with additional capnographic monitoring (“capnography arm”). Patients randomized to the standard arm were defined as controls. In both groups standard monitoring included clinical observation, pulse oximetry, automated blood pressure monitoring, and electrocardiography (ECG). In the capnography arm, capnographic data from the patients were available for additional noninvasive assessment of ventilation. Capnographic data of patients assigned to the standard arm were kept from view by folding an opaque cover over the capnography display, so that only the integrated pulse oximetric readout of the monitor was visible.
Monitoring and sedation regime during ERCP procedure All patients were monitored for clinical signs of respiratory abnormalities. Heart rate, pulse oximetry, and ECG changes were continuously assessed. Blood pressure was automatically noninvasively assessed at 3-minute intervals. Abnormal events detected on pulse oximetry, pulse rate, or blood pressure measurements were crosschecked for any mechanical issues related to devices and sensors. For assessment of oxygen saturation, the pulse oximeter integrated in the capnography device and an additional pulse oximeter were employed for double readings. A nasal cannula with an oral sampling port to accommodate mouth-breathers provided 2 L/min of oxygen and continuously sampled carbon dioxide (CO2) content of both inhaled and exhaled patient gas (Smart CapnoLine Guardian; Oridion Medical). The Smart CapnoLine Guardian set is designed to sample oral and nasal CO2 and administer supplemental oxygen via nose and mouth, for patients who can wear a 60-Fr bite block, during upper endoscopy-type procedures. The sampling line was connected to a portable bedside monitor (Capnostream 20; Oridion Medical) which displayed the time-based CO2 graphic waveform, the
numerical CO2 partial pressure (mmHg), the derived respiratory rate, and the oxygen saturation by integrated pulse oximetry (Nellcor, Covidien, Boulder, Colorado, USA). The height, shape, and rhythm of the capnogram provided a real-time assessment of ventilatory function. During inspiration, samples essentially contain no CO2 in the presence of supplemental oxygen. During expiration, samples represent alveolar CO2 concentration with a small amount of gas from patient physiologic dead space that contains no CO2. In the case of alveolar hypoventilation, samples show reduced or no CO2. All patients had a run-in of several minutes before the first dose of medication was given which signaled the beginning of sedation. In both groups an independent observer clinically monitored the patients for respiratory activity. In addition, the capnogram waveform enabled visual assessment of ventilation in the capnography arm. Here, the capnographic criteria for apnea were the absence of exhaled CO2 for a time period of 15 seconds. Midazolam and propofol were administered by a second physician with experience in intensive care medicine. The physician responsible for sedation was not involved in carrying out the endoscopy procedure, and all vital signs, including blood pressure, oxygen saturation, and pulse rate, were checked before administration of any sedation medication. Patients received a standard dose of 2.5 mg midazolam at the beginning of the procedure. Additionally, a loading dose of propofol (Propofol 1 %; Fresenius Kabi, Bad Homburg, Germany) adjusted to patient weight (40 mg at < 70 kg bodyweight or 60 mg at ≥ 70 kg bodyweight) was given immediately after the application of midazolam. In cases of advanced age or pre-existing severe co-morbidities, lower initial propofol doses were applied (e. g. 20 mg). Once the targeted level was reached, sedation was maintained by application of repeated doses of propofol (10 – 20 mg). We aimed to conduct ERCP under deep sedation. The level of sedation was monitored clinically by the physician responsible for sedation. Any sign of apnea or oxygen desaturation prompted an intervention which consisted of: (i) patient stimulation; (ii) withholding medication; (iii) a chin lift or jaw thrust maneuver; (iv) increasing oxygen supplementation; and (iv) insertion of a nasopharyngeal tube. We recorded indication, duration of procedure, and outcome of the investigation. The time of sedation was defined as the time between administration of the first dose and withdrawal of the endoscope from the mouth. At the end of the procedure, patient cooperation during endoscopy was rated by the endoscopist on a numeric analogue scale (NAS; 1 minimum to 10 maximum). Following completion of ERCP, patients were disconnected from electronic monitoring if they could give a meaningful verbal response and their vital sign parameters were stable. The patients were then transferred to the recovery area. Once they had attained full recovery patients were asked to rate their satisfaction with the sedation on an NAS (1 minimum to 10 maximum) and were then discharged from the endoscopy unit.
Randomization Randomization took place at the endoscopy unit of each participating site. The randomization sequence was created using nQuery Advisor version 7.0 with a prespecified block size. For allocation to either control or intervention group sealed envelopes were opened and randomization cards were retrieved. The endoscopy team did not take part in the randomization allocation process.
Klare Peter et al. Capnographic monitoring of midazolam and propofol sedation during ERCP… Endoscopy 2016; 48: 42–50
43
Downloaded by: University of Pittsburgh. Copyrighted material.
Original article
Original article
Assessed for eligibility n = 263 Excluded: n = 21 Eligibility criteria not met: n = 4 ▪Orotracheal intubation: n = 3 ▪Age below 18 years: n = 1 Refused participation: n = 11 Informed consent not possible: n = 6 ▪Language barrier: n = 2 ▪Legal guardianship: n = 4
Fig. 1 Capnographic monitoring of midazolam and propofol sedation during endoscopic retrograde cholangiopancreatography (ERCP): patient flow through EUS, endoscopic ultrasonography; PTBD, percutaneous transhepatic biliary drainage.
Enrolment n = 242
Randomization
Allocated to capnography arm n = 122 Received allocated intervention: n = 121 Did not receive allocated intervention: n=1 ▪ Procedure cancelled after randomization: n = 1 (Received EUS, 1)
Intention-to-treat analysis: n = 121 Per-protocol analysis: n = 108 Excluded from analysis because of protocol violation: n = 13 ▪ Received no oxygen supplementation: n=1 ▪ Received 4 L/min oxygen: n = 1 ▪ Persistent disruption of capnography: n = 11
Allocation
Analysis
Allocated to standard arm n = 120 Received allocated intervention: n = 117 Did not receive allocated intervention: n=3 ▪ Procedure cancelled after randomization: n = 3 (Received EUS, 2; received PTBD, 1)
Intention-to-treat analysis: n = 117 Per-protocol analysis: n = 115 Excluded from analysis because of protocol violation: n = 2 ▪ Received etomidate instead of propofol: n = 1 ▪ Received no oxygen supplementation: n=1
Outcome measures The primary outcome measure was the incidence of hypoxemic events, defined as a fall of Sao2 to < 90 %. Oxygen saturation was checked for hypoxemic events during the entire sedation period. Secondary outcome measures were: (i) apnea; (ii) further vital sign parameters, such as hypotension (systolic blood pressure < 90 mmHg) and bradycardia (heart rate < 50 /min); (iii) procedural parameters, in particular sedative dosage and duration of procedure; (iv) patient satisfaction; and (v) patient cooperation as rated by the endoscopist.
test for two independent samples or the Mann-Whitney U test were used, depending on the distribution of the data. We performed analyses both according to the intention-to-treat principle and per-protocol. A two-tailed P value of < 0.05 was considered statistically significant. Odds ratios (OR) and relative frequencies are presented with 95 % confidence intervals (95 %CI). Statistical analysis was done using statistical software R version 3.1.1 [6] and StatXact version10 (Cytel Studio Version 10.0, Cytel Software Corporation).
Sample size calculation
Results
We assumed that hypoxemic events occur in 30 % of all cases (of ERCP under sedation). An absolute difference of 15 percentage points was considered to be clinically relevant, i. e., a reduction to a hypoxemic event rate of 15 %. We applied the use of the χ² test for analysis of the primary end point. Under the given assumptions, a sample size of 121 cases was required for each group in order to detect a difference of 15 percentage points in hypoxemic events between both study groups with a power of 80 % using a two-sided significance rate of 5 %. Thus, a sample size of 242 patients was calculated.
!
Statistical analysis Continuous data are described by mean (standard deviation [SD]) or median (range); for categorical data absolute and relative frequencies are presented. Group comparisons of categorical data, including the primary end point of frequency of hypoxemic events, were conducted using χ² tests. If the number of expected cell counts was smaller than 5 for more than 20 % of the cells, Fisher’s exact test was used. For quantitative data, Student’s t
Intention-to-treat analysis A total of 263 patients were screened for enrolment. In 4 cases, patients failed to meet eligibility criteria; 11 patients refused to participate; and in 6 other cases written informed consent was not possible. Thus, 122 and 120 patients, respectively, were randomized to the capnography arm and standard arm. In the capnography arm 1 patient had to be excluded because endoscopic ultrasonography (EUS) was done instead of ERCP. In the standard arm 2 patients received EUS and 1 received percutaneous transhepatic biliary drainage (PTBD) instead of ERCP, which led to exclusion. Thus, 238 patients were included in the intention-to" Fig. 1). treat analysis (●
Patient, clinical and procedural characteristics: intention-to-treat Patient characteristics were well balanced with regard to age, gender, body composition and medical history with all of the ex" Table 1, P values not ploratory two-sided P values being > 0.05 (●
Klare Peter et al. Capnographic monitoring of midazolam and propofol sedation during ERCP… Endoscopy 2016; 48: 42–50
Downloaded by: University of Pittsburgh. Copyrighted material.
44
Original article
Factor
Capnography arm
Standard arm
( n = 121)
(n = 117)
Age, mean (SD), years
62.2 (14.2)
Male sex, n (%)
70 (57.9)
Body mass index, mean (SD), kg/m 2
61.7 (14.9) 57 (48.7)
25.1 (3.8)
25.5 (4.7)
Smoking (current and previous), n (%)
57 (54.8)
Alcohol abuse, current and previous, n (%)
14 (11.6)
56 (47.9) 9 (7.7)
Regular narcotic/sedative use, n (%)
18 (14.9)
14 (12.0)
Heart disease, n (%)
37 (30.6)
26 (22.2)
Lung disease, n (%)
13 (10.7)
10 (8.5)
Renal disease, n (%)
11 (9.17)
10 (8.5)
Liver disease, n (%)
38 (31.4)
35 (29.9)
Sleep apnea, n (%)
1 (0.8)
45
Table 1 Intention-to-treat analysis of monitoring with additional capnography versus standard monitoring of midazolam and propofol sedation during endoscopic retrograde cholangiopancreatography (ERCP): demographic, clinical, and procedural characteristics for patients. Data were available from all patients except where noted
2 (1.7) (Data from 116/117, 99.1 %)
I
19 (15.7)
21 (17.9)
II
53 (43.8)
53 (45.3)
III
45 (37.2)
37 (31.6)
IV
4 (3.3)
6 (5.1)
Mallampati class, n (%)
(Data from 119/121, 98.3 %)
(Data from 114/117, 97.4 %)
I
55 (46.2)
45 (39.5)
II
36 (30.3)
32 (28.1)
III
16 (13.4)
21 (18.4)
IV
12 (10.1)
16 (14.0)
Sedation problems during previous endoscopies, n (%)
3 (2.5)
5 (4.3)
Outpatient or ambulatory patients, n (%)
8 (6.6)
12 (10.3)
Benign biliary stenosis
33 (27.3)
21 (17.9)
Malignant biliary stenosis
37 (30.6)
33 (28.2)
Choledocholithiasis
22 (18.2)
31 (26.5)
Stent extraction
7 (5.8)
10 (8.5)
Biliary leaks
4 (3.3)
7 (6.0)
Indications, n (%)
Chronic pancreatitis
13 (10.7)
6 (5.1)
5 (4.1)
9 (7.7)
Cholangioscopy, n (%)
6 (5.0)
1 (0.9)
Altered anatomy, n (%)
1 (0.8)
5 (4.3)
Others
Baseline oxygen saturation, mean (SD), %
98.6 (1.5)
98.6 (1.6)
Baseline heart rate, mean (SD), beats/min
77.6 (13.7)
80.2 (16.1)
Baseline systolic blood pressure, mean (SD), mmHg
143.5 (21.8)
143.9 (21.7)
Total propofol dose, median (range), mg
400.0 (110 – 1610)
390.0 (80 – 1160) (Data from 116/117, 99.1 %)
38.0 (6 – 165)
Procedure time, median (range), min
shown). Malignant biliary stenosis was the most common indication for ERCP in both study arms. Baseline demographic, clinical, and procedural characteristics for both groups are presented in ●" Table 1.
Primary study outcome: intention-to-treat At least one episode of hypoxemia (Sao2 < 90 %) occurred in 46 patients in the capnography arm (38.0 %, 95 %CI 29.4 % – 46.9 %]) and in 52 patients in the standard arm (44.4 %, 95 %CI 35.3 % – 53.9 %) (P = 0.314 for difference in relative frequencies; OR = 1.3, 95 %CI " Table 2). In the capnography arm a total of 143 hy0.78 – 2.19; ●
38.0 (5 – 164) (Data from 116/117, 99.1 %)
poxemic events were observed compared with 158 hypoxemic events in the standard arm.
Secondary study outcomes: intention-to-treat Vital sign parameters. Apnea was detected in 78 patients in the capnography arm (64.5 %, 95 %CI 55.6 % – 73.0 %) and in 7 patients in the standard arm (6.0 %, 95 %CI 2.8 % – 11.5 %) (P < 0.001 for difference). In the capnography arm a total number of 284 events of apnea were detected. The capnography device indicated another 76 events which did not correspond to apnea clinically based upon the assessment of the independent observer. Technical problems accounted for 34
Klare Peter et al. Capnographic monitoring of midazolam and propofol sedation during ERCP… Endoscopy 2016; 48: 42–50
Downloaded by: University of Pittsburgh. Copyrighted material.
ASA class, n (%)
Original article
Table 2 Intention-to-treat analysis of prespecified outcomes in patient groups with additional capnographic monitoring or with standard monitoring of midazolam and propofol sedation during ERCP. Data were available for all patients except when noted otherwise.
Factor
Capnography arm (n = 121)
Standard arm (n = 117)
46 (38.0 %)
52 (44.4 %)
P value1
Primary end point, n (%) Oxygen saturation < 90 %
0.314
Secondary end points, n (%) Detection of apnea
78 (64.5 %)
7 (6.0 %)
Oxygen saturation ≤ 85 %
27 (22.3 %)
31 (26.5 %)
0.453
Oxygen saturation < 80 %
15 (12.4 %)
15 (12.8 %)
0.922
Oxygen supplementation increased
38 (31.4 %)
42 (35.9 %)
0.463
1 (0.8 %)
1 (0.9 %)
0.981
Assisted ventilation Death
< 0.001
0 (0.0 %)
1 (0.9 %)
0.308
Bradycardia
11 (9.1 %)
11 (8.5 %)
0.934
Hypotension
6 (5.0 %)
12 (10.3 %)
0.122
Patient cooperation rated by the endoscopist (NAS, 1 – 10), median (range)
8.0 (2 – 10) (Data available: 118/121, 97.5 %)
8.0 (1 – 10) (Data available: 113/117, 96.6 %)
0.271
Patient satisfaction (NAS, 1 – 10), median (range)
9.0 (2 – 10) (Data available: 95/121, 78.5 %)
9.5 (3 – 10) (Data available: 90/117, 76.9 %)
1
To eye-opening
5.0 (0 – 21) (Data available: 103/121, 85.1 %)
5.0 (1 – 21) (Data available: 101/117, 86.3 %)
0.841
To leaving of procedure room
7.0 (1 – 23) (Data available: 102/121, 84.3 %)
7.0 (2 – 21) (Data available: 93/117, 79.5 %)
0.689
Time from end of endoscopy, median (range) min
NAS, numerical analogue scale 1 P values derived from Student t tests, Mann-Whitney U tests, and Fisher`s exact tests as appropriate.
of these 76 false alarms (44.7 %), namely: dislocation of the nasal cannula, 11 events; blocking of the filter line, 10 events; orotracheal suctioning, 5 cases; and insertion of a nasopharyngeal tube, 8 cases. In the remaining 42 events (55.3 %) no specific cause could be identified. Severe hypoxemia (defined as Sao2 < 85 %) occurred in 27 patients (22.3 %) in the capnography arm and in 31 patients (26.5 %) in the standard arm (P = 0.453 for difference; OR = 1.26, 95 %CI 0.69 – 2.27). There were no differences between the two groups regarding the incidence rates of increased oxygen supplementation (P = 0.463), bradycardia (P = 0.934), and hypoten" Table 2). sion (P = 0.122) (● Patient cooperation, quality of sedation, and recovery. Cooperation scores as rated by the endoscopists after ERCP procedures did not differ between the study arms (8.0 [2 – 10] vs. 8.0 [1 – 10]; P = 0.271). Patient satisfaction was high in both groups and no difference could be detected between the capnography arm and the standard arm (9.0 [2 – 10] vs. 9.5 [3 – 10]; P = 1.000). The use of capnography had no influence on the time from end of endoscopy to eye-opening (5.0 min [0 – 21]) vs. 5.0 min [1 – 21]; P = 0.841) or to transfer of the patient to the recovery area (capnography 7.0 min [1 – 23] vs. standard 7.0 min [2 – 21]; P = 0.689).
guidelines. Additionally, intravenous thrombolysis was administered because of the possibility of an acute thromboembolic event. Under high catecholamine doses circulation was restored approximately 45 minutes after the beginning of cardiopulmonary resuscitation. The patient was then transferred to the intensive care unit. Unfortunately, the patient died shortly thereafter because of repeated circulatory arrest. Unblinded capnography data including end-tidal CO2 values were analyzed retrospectively. Data indicated that hypoxemia and apnea occurred simultaneously 52 minutes after the beginning of ERCP. During the following minutes of the procedure capnography showed hypoventilation. Shortly before the development of critical hypotension, bradycardia and hypoxemia end-tidal CO2 values indicated another episode of apnea. The abovementioned death was rated as a serious adverse event and was therefore reported to the Ethics Committee of the Technische Universität München, Munich, Germany. The study was stopped until written evaluation of the incident by the ethics committee was available. The committee found that the incident was not triggered by the participation in the study itself. We therefore continued inclusion in May 2011.
Per-protocol analysis Serious adverse events In April 2011, one patient in the standard arm developed severe hypoxemia approximately 52 minutes after the beginning of the ERCP procedure. Up to this point a total dose of 2.5 mg midazolam and 300 mg propofol had been given. Administration of sedatives was stopped immediately and oxygen supplementation was increased. At 6 minutes after the first event of oxygen desaturation the patient developed critical hypotension (systolic blood pressure < 90 mmHg) and bradycardia (heart rate < 50/min). Cardiopulmonary arrest then occurred. Cardiopulmonary resuscitation was started immediately and orotracheal intubation was conducted. During the following 45 minutes of resuscitation, adrenaline was given repeatedly according to current resuscitation
Of the 238 patients, who were included in the intention-to-treat analysis, 15 were excluded from the per-protocol analysis. Among these, 2 accidentally received no oxygen supplementation; 1 other patient received 4 L/min oxygen during the procedure; and 1 patient received etomidate instead of propofol. In 11 cases in the capnography arm, continuing problems using the capnography device were reported that made the performance of capnography impossible. These cases were rated as protocol violations. All 11 cases were caused by practical problems in applying capnography: continuing ongoing problems placing the sample line in 6 cases; need to exchange the oxygen probes in 3 cases; and blocked filter lines in 2 cases. Thus, 223 patients were included in the per-protocol analysis.
Klare Peter et al. Capnographic monitoring of midazolam and propofol sedation during ERCP… Endoscopy 2016; 48: 42–50
Downloaded by: University of Pittsburgh. Copyrighted material.
46
Original article
Factor
Capnography arm
Standard arm
(n = 108)
(n = 115)
Age, mean (SD), years
62.1 (14.1)
61.9 (14.8)
Male sex, n (%)
60 (55.6)
57 (49.6)
Body mass index, mean (SD), kg/m 2
24.8 (3.7)
25.5 (4.7)
Smoking (current and previous), n (%)
50 (46.3)
56 (48.7)
Alcohol abuse (current and previous), n (%)
13 (12.0)
9 (7.8)
Regular narcotic/sedative use, n (%)
18 (16.7)
14 (12.2)
Heart disease, n (%)
29 (26.9)
26 (22.6)
Lung disease, n (%)
12 (11.1)
10 (8.7)
Renal disease, n (%)
11 (10.2)
10 (8.7)
Liver disease, n (%)
34 (31.5)
34 (29.6)
Sleep apnea, n (%)
1 (0.9)
47
Table 3 Per-protocol analysis of monitoring with additional capnographic monitoring versus standard monitoring of midazolam and propofol sedation during ERCP: demographic, clinical, and procedural characteristics for patients. Data available from all patients except where noted.
1 (0.9), (Data from 114 /115, 99.1 %)
I
17 (15.7)
20 (17.4)
II
49 (45.4)
53 (46.1)
III
39 (36.1)
36 (31.3)
IV
3 (2.8)
6 (5.2)
Mallampati class, n (%)
(Data from 106 /108, 98.1 %)
(Data from 112 /115, 97.4 %)
I
52 (49.1)
45 (40.2)
II
30 (28.3)
31 (27.7)
III
13 (12.3)
21 (18.6)
IV
11 (10.4)
15 (13.4)
Sedation problems during previous endoscopies, n (%)
3 (2.8)
5 (4.3)
Outpatient or ambulatory patients, n (%)
8 (7.4)
11 (9.6)
Benign biliary stenosis
31 (28.7)
21 (18.3)
Malignant biliary stenosis
32 (29.6)
32 (27.8)
Choledocholithiasis
18 (16.7)
31 (27.0)
Stent extraction
7 (6.5)
9 (7.8)
Biliary leaks
4 (3.7)
7 (6.1)
11 (10.2)
6 (5.2)
5 (4.6)
9 (7.8)
Indications, n (%)
Chronic pancreatitis Others Cholangioscopy
6 (5.6)
1 (0.9)
Altered anatomy
1 (0.9)
5 (4.3)
Baseline oxygen saturation, mean (SD), %
98.6 (1.5)
98.6 (1.6)
Baseline heart rate mean (SD), beats/min
78.2 (13.6)
80.2 (16.2)
Baseline systolic blood pressure, mean (SD), mmHg
144.3 (20.5)
144.0 (21.7)
Total propofol dose, median (range), mg
390.0 (110 – 1610)
380.0 (80 – 1660)
37.0 (6 – 165)
Procedure time, median (range), min
38.0 (5 – 164) (Data from 114 /115, 99.1 %)
P values derived from to Student t tests, Mann-Whitney U tests and Pearson chi-squared tests as appropriate.
Patient, clinical and procedural characteristics: per-protocol Baseline demographic, clinical, and procedural characteristics for " Table 3 and showed an equitable both groups are presented in ● distribution between the groups with all of the exploratory twosided P values being > 0.05 (P values not shown).
Primary study outcome: per-protocol Hypoxemia (at least one episode of Sao2 < 90 %) was observed in 34 patients in the capnography arm (31.5 %, 95 %CI 22.9 – 41.1 %) and 51 patients in the standard arm (44.4 %, 95 %CI 35.1 – 53.5 %) " Table 4). (P = 0.048 for difference; OR = 1.73, 95 %CI 1.003 – 3) (●
Secondary study outcomes: per-protocol Vital sign parameters. Apnea was detected in 66 patients in the capnography arm (61.1 %, 95 %CI 51.6 % – 70.3 %) and in 6 in the standard arm (5.2 %, 95 %CI 2.3 % – 11 %) (P < 0.001 for difference). Severe hypoxemia (defined as Sao2 < 85 %) was observed less frequently in the capnography arm compared with the standard arm (16 [14.8 %] vs. 30 [26.1 %], P = 0.038; OR = 2.03, 95 %CI 1.03 – 3.98). The incidence of increased oxygen supplementation was lower in the capnography arm compared with controls (24 [22.2 %] vs. 41 [35.7 %], P = 0.028; OR = 1.94, 95 %CI 1.07 – 3.5]). There were no differences between incidence rates of bradycar-
Klare Peter et al. Capnographic monitoring of midazolam and propofol sedation during ERCP… Endoscopy 2016; 48: 42–50
Downloaded by: University of Pittsburgh. Copyrighted material.
ASA class, n (%)
Original article
Table 4 Per-protocol analysis of prespecified outcomes in patient groups with additional capnographic monitoring or with standard monitoring of midazolam and propofol sedation during ERCP. Data were available for all patients unless otherwise noted.
Factor
Capnography arm (n = 108)
Standard arm (n = 115)
34 (31.5 %)
51 (44.4 %)
P value1
Primary end point, n (%) Oxygen saturation < 90 %
0.048
Secondary end points, n (%) Detection of apnea
66 (61.1 %)
6 (5.2 %)
Oxygen saturation ≤ 85 %
16 (14.8 %)
30 (26.1 %)
0.038
Oxygen saturation < 80 %
7 (6.5 %)
14 (12.2 %)
0.146 0.028
Oxygen supplementation increased
< 0.001
24 (22.2 %)
41 (35.7 %)
Assisted ventilation
0 (0.0 %)
1 (0.9 %)
Death
0 (0.0 %)
1 (0.9 %)
Bradycardia
9 (8.3 %)
11 (10.2 %)
0.748
Hypotension
5 (4.6 %)
1 1
12 (10.4 %)
0.103
Patient cooperation rated by the endoscopist (NAS, 1 – 10), median (range)
8.0 (2 – 10) (Data available: 106 /108, 98.1 %)
8.0 (1 – 10) (Data available: 111 /115, 96.2 %)
0.082
Patient satisfaction (NAS, 1 – 10), median (range)
9.0 (2 – 10) (Data available: 88 /108 (81.5 %)
9.0 (3 – 10) (Data available: 89 /115 (77.4 %)
0.992
To eye-opening
5.9 (0 – 21) (Data available: 92 /108 (85.2 %)
5.6 (1 – 21) (Data available: 99 /115 (86.1 %)
0.944
To leaving of procedure room
7.8 (1 – 23) (Data available: 91 /108 (84.3 %)
7.5 (2 – 21) (Data available: 91 /115 (79.1 %)
0.944
Time from end of endoscopy, median (range) min
NAS, numerical analogue scale 1 P values derived from Student t tests, Fisher`s exact tests, and Mann-Whitney U tests, as appropriate.
dia (P = 0.748) and hypotension (P = 0.103) between the two " Table 4). groups (● Patient cooperation, quality of sedation, and recovery. Cooperation scores as rated by the endoscopist after ERCP procedures did not differ between the study groups (capnography 8.0 [2 – 10] vs. standard 8.0 [1 – 10], P = 0.082). Patient satisfaction scores were equal in the two groups (9.0 [2 – 10] vs. 9.0 [3 – 10]; P = 0.992). The use of capnography had no influence on the time from end of endoscopy to eye-opening (capnography 5.9 min [0 – 21] vs. standard 5.6 min [1 – 21], P = 0.944) or to transfer of the patient to the recovery area (capnography 7.8 min [1 – 23] vs. standard 7.5 min [2 – 21], P = 0.944).
Discussion !
The use of sedatives has improved patient comfort and acceptance with regard to endoscopy in recent years [7]. For invasive and potentially painful investigations especially, sedation is mandatory. Propofol is widely used for endoscopy but has the potential to cause cardiorespiratory side effects [8]. Capnography monitoring can detect apnea in the nonintubated patient [9], but nevertheless, the use of capnography is not established as a recognized standard for monitoring respiratory function during endoscopy [7, 10 – 13].
The influence of capnography on incidence of hypoxemic events In this study we aimed to show that capnography-directed early interventions are useful in avoiding sedation-induced hypoxemia. We found that hypoxemia incidence rates did not differ significantly between both groups in an intention-to-treat analysis. However, capnography was superior to standard oximetry in detecting apnea episodes. Capnography monitoring has been shown to help avoid desaturation in standard endoscopy procedures, such as esophagoduode-
noscopy (EGD) and colonoscopy [14, 15]. Moreover, in one trial comparing capnography versus oximetry monitoring in patients undergoing upper endosonography or ERCP, the frequency of hypoxemia was significantly reduced in the capnography arm [16] which, however, contradicts our results. However, it must be noted that in the aforementioned trial benzodiazepines and opioids were used for sedation and most of the patients (approximately 75 %) received endosonography, a less invasive investigation than ERCP [16]. One explanation for our results could be that conditions related to ERCP itself hampered effective interventions, even though good and ample attention was given to apnea events. Patients were positioned in a prone position which might act as a limiting factor when rescue maneuvers (e. g. chin-lift) are being used. We therefore assume that interventions based on the detection of apnea might have been ineffective in some cases. On the other hand we experienced some technical and practical problems which interfered with capnography measurement. Technical difficulties with capnography sampling have also been described previously [15]. In our study, we detected a remarkable number of false-positive apnea alarms given by capnography which were mainly due to technical issues such as dislocation of the nasal cannula. Difficult or invasive interventions during ERCP may lead to pain or discomfort which consequently result in agitation. This may be one out of several reasons why capnography measuring was error-prone in our trial. Stable measuring conditions for the duration of the whole procedure would be a prerequisite for the correct detection of apnea and, in consequence, for efficiently conducted countermeasures to avoid hypoxemia. These hypotheses are further supported by the results of the perprotocol analysis. When cases with persistent mismeasurement were excluded, we found a lower incidence of hypoxemia in the capnography arm compared with the standard arm. Furthermore, severe hypoxemic events also occurred less frequently in patients in the capnography arm. One possible interpretation might be that if capnography could be applied correctly without
Klare Peter et al. Capnographic monitoring of midazolam and propofol sedation during ERCP… Endoscopy 2016; 48: 42–50
Downloaded by: University of Pittsburgh. Copyrighted material.
48
Original article
Serious adverse event during the trial Unfortunately, in one patient receiving standard monitoring cardiopulmonary arrest occurred during ERCP, which subsequently led to the death of the patient. The exact cause of the event remained unknown although cardiorespiratory depression related to sedation seemed probable. The unblinded capnography data retrospectively showed that apnea episodes occurred immediately prior to the development of hypoxemia and cardiac arrest in this case. The preventability of this event remains a subject of speculation. Sedation-related death during endoscopy procedures has been shown to occur rarely [17]. Furthermore, high patient morbidity (ASA class III and higher) as well as emergency investigations are factors which predict an increased risk for severe events [18]. In our case the patient was classified as ASA II, and was suffering from biliodigestive stenosis after surgery (Whipple procedure) for carcinoma of the papilla of vater. In summary this tragic incident shows that endoscopic procedures under sedation bear a significant risk even in patients who present in a relatively good general condition.
Study limitations Our study is subject to several limitations. Neither patients nor the endoscopist and physician responsible for sedation nor the independent observer were blinded to assignment of patients. Thus there was the potential for bias regarding the assessment of apnea episodes. Secondly, as no validated scale was used to assess depth of sedation, this could have differed between the two groups and could have led to differences in the incidence of oxygen desaturation and hypoxemia. However, the mean total propofol doses did not differ between the groups, suggesting that there was no bias regarding the application of propofol. Furthermore, hypoxemic events were observed frequently without the occurrence of clinically relevant consequences. This illustrates that hypoxemia only served as a surrogate parameter. The occurrence of hypoxemia might not be interpreted as a sedationrelated adverse event in itself. One way of addressing this issue could be to primarily focus on severe hypoxemia (Sao2 < 85 %). However, it is unclear whether severe hypoxemia is more feasible in serving as an adequate surrogate for sedation-related complications. Furthermore, any fall of Sao2 < 90 % is considered to be important, as stated by gastroenterological associations [7, 11]. We therefore decided to use the latter threshold as target indicator in our trial. Finally, oxygen supplementation may have masked hypoventilation, leading to artificially high Sao2 measurements. However, current national guidelines recommend routine use of oxygen supplementation for procedural sedation during endoscopy [7].
Conclusion In summary, our study showed, that capnography can adequately detect apnea in the setting of ERCP under midazolam and propofol sedation. Based on the results of the intention-to-treat analysis, no beneficial effect of capnography-based early intervention could be demonstrated. However, capnographic measurement
was hindered in several patients because of particular circumstances during ERCP. After the exclusion of cases with technical problems, per-protocol analysis revealed a lower incidence of hypoxemia in the capnography arm compared with the standard monitoring arm. We therefore conclude that the practical application of capnography needs to be improved. If this is possible, capnography may be of value in improving patient safety in the future. In this case, the results of our study should be noted when selecting monitoring devices for routine ERCP procedures. Competing interests: Stefan von Delius, Andrea Riphaus and Till Wehrmann received material support for research purposes from Oridion Medical (Needham, Massachusetts, U.S.). Oridion Medical provided the capnography monitor (Capnostream 20) and sample lines (Smart CapnoLine Guardian), but did not participate in study design, data collection, data analysis, or manuscript preparation. No financial support was received for this study. Institutions 1 II. Medizinische Klinik und Poliklinik, Klinikum rechts der Isar der Technischen Universität München, Germany 2 Klinik für Innere Medizin I, Universitätsklinikum Ulm, Germany 3 Institut für Medizinische Biometrie, Epidemiologie und Medizinische Informatik, Universität des Saarlandes, Campus Homburg, Germany 4 Medizinische Universitätsklinik des Knappschaftskrankenhaus der Ruhr-Universität Bochum, Germany 5 Klinik für Anästhesie, Universitätsklinikum der Universität Witten/Herdecke, Helios Klinikum Wuppertal, Germany 6 Fachbereich Gastroenterologie, Deutsche Klinik für Diagnostik, Helios Klinik Wiesbaden, Germany 7 Medizinische Klinik, KRH Klinikum Agnes Karll, Laatzen, Germany
References 1 McQuaid KR, Laine L. A systematic review and meta-analysis of randomized, controlled trials of moderate sedation for routine endoscopic procedures. Gastrointest Endosc 2008; 67: 910 – 923 2 Gerstenberger PD. Capnography and patient safety for endoscopy. Clin Gastroenterol Hepatol 2010; 8: 423 – 425 3 Vargo JJ, Zuccaro GJr, Dumot JA et al. Automated graphic assessment of respiratory activity is superior to pulse oximetry and visual assessment for the detection of early respiratory depression during therapeutic upper endoscopy. Gastrointest Endosc 2002; 55: 826 – 831 4 Burton JH, Harrah JD, Germann CA et al. Does end-tidal carbon dioxide monitoring detect respiratory events prior to current sedation monitoring practices? Acad Emerg Med 2006; 13: 500 – 504 5 Deitch K, Miner J, Chudnofsky CR et al. Does end tidal CO2 monitoring during emergency department procedural sedation and analgesia with propofol decrease the incidence of hypoxic events? A randomized, controlled trial Ann Emerg Med 2010; 55: 258 – 264 6 R Core Team. R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing; 2014: http://www.Rproject.org/ 7 Riphaus A, Wehrmann T, Weber B et al. [S3-guidelines – sedation in gastrointestinal endoscopy]. Z Gastroenterol 2008; 46: 1298 – 1330 DOI 10.1055/s-2008-1027850 8 Bell GD. Preparation, premedication, and surveillance. Endoscopy 2004; 36: 23 – 31 9 Manifold CA, Davids N, Villers LC et al. Capnography for the nonintubated patient in the emergency setting. J Emerg Med 2013; 45: 626 – 632 10 Lichtenstein DR, Jagannath S, Baron TH et al. Sedation and anesthesia in GI endoscopy. Gastrointest Endosc 2008; 68: 815 – 826 11 Dumonceau JM, Riphaus A, Aparicio JR et al. European Society of Gastrointestinal Endoscopy, European Society of Gastroenterology and Endoscopy Nurses and Associates, and the European Society of Anaesthesiology Guideline: Non-anesthesiologist administration of propofol for GI endoscopy. Endoscopy 2010; 42: 960 – 974 12 ASGE Technology Committee Gottlieb KT, Banerjee S, Barth BA et al. Monitoring equipment for endoscopy. Gastrointest Endosc 2013; 77: 175 – 180
Klare Peter et al. Capnographic monitoring of midazolam and propofol sedation during ERCP… Endoscopy 2016; 48: 42–50
Downloaded by: University of Pittsburgh. Copyrighted material.
any disruption, this can avoid hypoxemia in patients receiving ERCP under propofol and midazolam sedation. We therefore conclude that efforts should be directed to improve the application of capnography in ERCP settings. In particular, there is a need to improve sampling under difficult conditions. If this were possible, further studies would be welcome to verify our assumption.
49
Original article
13 American Society for Gastrointestinal Endoscopy, American Gastroenterological Association, American College of Gastroenterology. MultiSociety Statement: Universal adoption of capnography for moderate sedation in adults undergoing upper endoscopy and colonoscopy has not been shown to improve patient safety or clinical outcomes and significantly increases costs for moderate sedation. http://www.asge.org/ WorkArea/showcontent.aspx?id=14226 [Accessed: January 10, 2015] 14 Lightdale JR, Goldmann DA, Feldman HA et al. Microstream capnography improves patient monitoring during moderate sedation: a randomized, controlled trial. Pediatrics 2006; 117: e1170 – e1178 15 Beitz A, Riphaus A, Meining A et al. Capnographic monitoring reduces the incidence of arterial oxygen desaturation and hypoxemia during propofol sedation for colonoscopy: a randomized, controlled study (ColoCap Study). Am J Gastroenterol 2012; 107: 1205 – 1212
16 Qadeer MA, Vargo JJ, Dumot JA et al. Capnographic monitoring of respiratory activity improves safety of sedation for endoscopic cholangiopancreatography and ultrasonography. Gastroenterology 2009; 136: 1568 – 1576 17 Rex DK, Deenadayalu VP, Eid E et al. Endoscopist-directed administration of propofol: a worldwide safety experience. Gastroenterology 2009; 137: 1229 – 1237 18 Behrens A, Labenz J, Schuler A et al. [How safe is sedation in gastrointestinal endoscopy? A multicentre analysis of 388,404 endoscopies and analysis of data from prospective registries of complications managed by members of the Working Group of Leading Hospital Gastroenterologists (ALGK)] Z Gastroenterol 2013; 51: 432 – 436
Downloaded by: University of Pittsburgh. Copyrighted material.
50
Klare Peter et al. Capnographic monitoring of midazolam and propofol sedation during ERCP… Endoscopy 2016; 48: 42–50
Capnographic monitoring of midazolam and propofol sedation during ERCP: a randomized controlled study (EndoBreath Study)
Authors
Peter Klare1, Johanna Reiter1, Alexander Meining2, Stefan Wagenpfeil3, Tim Kronshage4, Christoph Geist4, Stefan Heringlake4, Christoph Schlag1, Monther Bajbouj1, Gerhard Schneider5, Roland M. Schmid1, Till Wehrmann6, *, Stefan von Delius1, *, Andrea Riphaus7, *
Institutions
Institutions are listed at end of article.
submitted 11. April 2015 accepted after revision 13. July 2015
Background and study aims: This was to determine whether intervention based on additional capnographic monitoring reduces the incidence of hypoxemia during midazolam and propofol sedation for endoscopic retrograde cholangiopancreatography (ERCP). Methods: Patients (American Society of Anesthesiologists [ASA] I – IV) scheduled for ERCP under midazolam and propofol sedation were randomly assigned to a control arm with standard monitoring or an interventional arm with additional capnographic monitoring. In both arms detection of apnea prompted withholding of propofol administration, stimulation of the patient, insertion of a nasopharyngeal tube, or further measures. The primary study end point was incidence of hypoxemia (oxygen saturation [Sao2] below 90 %); secondary end points included occurrences of severe hypoxemia (Sao2 ≤ 85 %), bradycardia, and hypotension, and sedation quality (patient cooperation and satisfaction).
Bibliography DOI http://dx.doi.org/ 10.1055/s-0034-1393117 Published online: 28.9.2015 Endoscopy 2016; 48: 42–50 © Georg Thieme Verlag KG Stuttgart · New York ISSN 0013-726X Corresponding author: Stefan von Delius, MD Klinikum rechts der Isar der Technischen Universität München II. Medizinische Klinik Ismaninger Str. 22, 81675 Munich Germany Fax: +49-89-41404905 [email protected]
Results: 242 patients were enrolled at three German endoscopy centers. Intention-to-treat analysis revealed no significant reduction in hypoxemia incidence in the capnography arm compared with the standard arm (38.0 % vs. 44.4 %, P = 0.314). Apnea was more frequently detected in the capnography arm (64.5 % vs. 6.0 %, P < 0.001). There were no differences regarding rates of bradycardia and hypotension. Per-protocol analysis showed lower incidence of hypoxemia in the capnography arm compared with the standard arm (31.5 % vs. 44.8 %, P = 0.048). There was one death related to sedation in the standard arm. Sedation quality was similar in the two groups. Conclusion: Intention-to-treat analysis showed hypoxemia incidence was not significantly lower in the additional capnography arm compared with standard monitoring. Additional capnographic monitoring of ventilatory activity resulted in improved detection of apnea. ClinicalTrials.gov identifier: NCT01072474.
Introduction !
Endoscopic retrograde cholangiopancreatography is an invasive procedure that requires the use of sedation [1]. During sedation, cardiopulmonary complications such as apnea and consequent hypoxemia may occur and demand monitoring of the patient. Therefore, electronic monitoring as well as close observation of patient responsiveness and respiratory activity is mandatory [1 – 3]. However, neither clinical observation nor pulse oximetry enable early detection of hypoventilation, apnea, and hypercarbia, or their consequences, including acidosis, catecholamine release, myocardial excitation with arrhythmias, and myocardial depression [2]. On the other
hand, apnea and altered respiration frequently precede hypoxemia with a significant time lag [2, 3]. It has been shown that capnographic monitoring is superior to pulse oximetry and visual assessment for the detection of respiratory depression [4, 5]. Hence, capnography could improve patient monitoring and prevent sedation-related hypoxemia, morbidity, and mortality. The aim of this randomized controlled study was to determine whether early intervention based upon data from capnographic monitoring could reduce the incidence of hypoxemia during midazolam and propofol sedation for ERCP.
* These authors contributed equally to the study.
Klare Peter et al. Capnographic monitoring of midazolam and propofol sedation during ERCP… Endoscopy 2016; 48: 42–50
Downloaded by: University of Pittsburgh. Copyrighted material.
42
Patients and methods !
Study population The study protocol was approved by the local ethics committee and registered at ClinicalTrials.gov (identifier: NCT01072474). Patient enrolment began in February 2010 and concluded in October 2011. Outpatients and inpatients aged > 18 years who were scheduled for ERCP at one of the three participating study sites (II Medizinische Klinik und Poliklinik, Klinikum rechts der Isar der Technischen Universität München, Munich, Germany; Medizinische Universitätsklinik des Knappschaftskrankenhaus der Ruhr-Universität Bochum, Bochum, Germany; and Deutsche Klinik für Diagnostik, Helios Klinik Wiesbaden, Wiesbaden, Germany) were eligible for participation. Exclusion criteria were American Society of Anesthesiologists [ASA] risk classification V, pregnancy, or pre-existing hypoxemia (Sao2 < 90 %), hypotension (systolic blood pressure < 90 mmHg), or bradycardia (heart rate < 50 /min) before the start of the endoscopy. Written informed consent was obtained from all patients.
Study design We conducted a randomized controlled trial at three participating sites in Germany. We assessed the efficacy of early intervention prompted by additional capnographic monitoring (Capnostream 20; Oridion Medical, Needham, Massachusetts, USA) to reduce the incidence of arterial oxygen desaturation (defined as a fall to below 90 %) during ERCP under propofol and midazolam sedation. Patients were randomly assigned to an arm with standard monitoring (“standard arm”) and to an arm with additional capnographic monitoring (“capnography arm”). Patients randomized to the standard arm were defined as controls. In both groups standard monitoring included clinical observation, pulse oximetry, automated blood pressure monitoring, and electrocardiography (ECG). In the capnography arm, capnographic data from the patients were available for additional noninvasive assessment of ventilation. Capnographic data of patients assigned to the standard arm were kept from view by folding an opaque cover over the capnography display, so that only the integrated pulse oximetric readout of the monitor was visible.
Monitoring and sedation regime during ERCP procedure All patients were monitored for clinical signs of respiratory abnormalities. Heart rate, pulse oximetry, and ECG changes were continuously assessed. Blood pressure was automatically noninvasively assessed at 3-minute intervals. Abnormal events detected on pulse oximetry, pulse rate, or blood pressure measurements were crosschecked for any mechanical issues related to devices and sensors. For assessment of oxygen saturation, the pulse oximeter integrated in the capnography device and an additional pulse oximeter were employed for double readings. A nasal cannula with an oral sampling port to accommodate mouth-breathers provided 2 L/min of oxygen and continuously sampled carbon dioxide (CO2) content of both inhaled and exhaled patient gas (Smart CapnoLine Guardian; Oridion Medical). The Smart CapnoLine Guardian set is designed to sample oral and nasal CO2 and administer supplemental oxygen via nose and mouth, for patients who can wear a 60-Fr bite block, during upper endoscopy-type procedures. The sampling line was connected to a portable bedside monitor (Capnostream 20; Oridion Medical) which displayed the time-based CO2 graphic waveform, the
numerical CO2 partial pressure (mmHg), the derived respiratory rate, and the oxygen saturation by integrated pulse oximetry (Nellcor, Covidien, Boulder, Colorado, USA). The height, shape, and rhythm of the capnogram provided a real-time assessment of ventilatory function. During inspiration, samples essentially contain no CO2 in the presence of supplemental oxygen. During expiration, samples represent alveolar CO2 concentration with a small amount of gas from patient physiologic dead space that contains no CO2. In the case of alveolar hypoventilation, samples show reduced or no CO2. All patients had a run-in of several minutes before the first dose of medication was given which signaled the beginning of sedation. In both groups an independent observer clinically monitored the patients for respiratory activity. In addition, the capnogram waveform enabled visual assessment of ventilation in the capnography arm. Here, the capnographic criteria for apnea were the absence of exhaled CO2 for a time period of 15 seconds. Midazolam and propofol were administered by a second physician with experience in intensive care medicine. The physician responsible for sedation was not involved in carrying out the endoscopy procedure, and all vital signs, including blood pressure, oxygen saturation, and pulse rate, were checked before administration of any sedation medication. Patients received a standard dose of 2.5 mg midazolam at the beginning of the procedure. Additionally, a loading dose of propofol (Propofol 1 %; Fresenius Kabi, Bad Homburg, Germany) adjusted to patient weight (40 mg at < 70 kg bodyweight or 60 mg at ≥ 70 kg bodyweight) was given immediately after the application of midazolam. In cases of advanced age or pre-existing severe co-morbidities, lower initial propofol doses were applied (e. g. 20 mg). Once the targeted level was reached, sedation was maintained by application of repeated doses of propofol (10 – 20 mg). We aimed to conduct ERCP under deep sedation. The level of sedation was monitored clinically by the physician responsible for sedation. Any sign of apnea or oxygen desaturation prompted an intervention which consisted of: (i) patient stimulation; (ii) withholding medication; (iii) a chin lift or jaw thrust maneuver; (iv) increasing oxygen supplementation; and (iv) insertion of a nasopharyngeal tube. We recorded indication, duration of procedure, and outcome of the investigation. The time of sedation was defined as the time between administration of the first dose and withdrawal of the endoscope from the mouth. At the end of the procedure, patient cooperation during endoscopy was rated by the endoscopist on a numeric analogue scale (NAS; 1 minimum to 10 maximum). Following completion of ERCP, patients were disconnected from electronic monitoring if they could give a meaningful verbal response and their vital sign parameters were stable. The patients were then transferred to the recovery area. Once they had attained full recovery patients were asked to rate their satisfaction with the sedation on an NAS (1 minimum to 10 maximum) and were then discharged from the endoscopy unit.
Randomization Randomization took place at the endoscopy unit of each participating site. The randomization sequence was created using nQuery Advisor version 7.0 with a prespecified block size. For allocation to either control or intervention group sealed envelopes were opened and randomization cards were retrieved. The endoscopy team did not take part in the randomization allocation process.
Klare Peter et al. Capnographic monitoring of midazolam and propofol sedation during ERCP… Endoscopy 2016; 48: 42–50
43
Downloaded by: University of Pittsburgh. Copyrighted material.
Original article
Original article
Assessed for eligibility n = 263 Excluded: n = 21 Eligibility criteria not met: n = 4 ▪Orotracheal intubation: n = 3 ▪Age below 18 years: n = 1 Refused participation: n = 11 Informed consent not possible: n = 6 ▪Language barrier: n = 2 ▪Legal guardianship: n = 4
Fig. 1 Capnographic monitoring of midazolam and propofol sedation during endoscopic retrograde cholangiopancreatography (ERCP): patient flow through EUS, endoscopic ultrasonography; PTBD, percutaneous transhepatic biliary drainage.
Enrolment n = 242
Randomization
Allocated to capnography arm n = 122 Received allocated intervention: n = 121 Did not receive allocated intervention: n=1 ▪ Procedure cancelled after randomization: n = 1 (Received EUS, 1)
Intention-to-treat analysis: n = 121 Per-protocol analysis: n = 108 Excluded from analysis because of protocol violation: n = 13 ▪ Received no oxygen supplementation: n=1 ▪ Received 4 L/min oxygen: n = 1 ▪ Persistent disruption of capnography: n = 11
Allocation
Analysis
Allocated to standard arm n = 120 Received allocated intervention: n = 117 Did not receive allocated intervention: n=3 ▪ Procedure cancelled after randomization: n = 3 (Received EUS, 2; received PTBD, 1)
Intention-to-treat analysis: n = 117 Per-protocol analysis: n = 115 Excluded from analysis because of protocol violation: n = 2 ▪ Received etomidate instead of propofol: n = 1 ▪ Received no oxygen supplementation: n=1
Outcome measures The primary outcome measure was the incidence of hypoxemic events, defined as a fall of Sao2 to < 90 %. Oxygen saturation was checked for hypoxemic events during the entire sedation period. Secondary outcome measures were: (i) apnea; (ii) further vital sign parameters, such as hypotension (systolic blood pressure < 90 mmHg) and bradycardia (heart rate < 50 /min); (iii) procedural parameters, in particular sedative dosage and duration of procedure; (iv) patient satisfaction; and (v) patient cooperation as rated by the endoscopist.
test for two independent samples or the Mann-Whitney U test were used, depending on the distribution of the data. We performed analyses both according to the intention-to-treat principle and per-protocol. A two-tailed P value of < 0.05 was considered statistically significant. Odds ratios (OR) and relative frequencies are presented with 95 % confidence intervals (95 %CI). Statistical analysis was done using statistical software R version 3.1.1 [6] and StatXact version10 (Cytel Studio Version 10.0, Cytel Software Corporation).
Sample size calculation
Results
We assumed that hypoxemic events occur in 30 % of all cases (of ERCP under sedation). An absolute difference of 15 percentage points was considered to be clinically relevant, i. e., a reduction to a hypoxemic event rate of 15 %. We applied the use of the χ² test for analysis of the primary end point. Under the given assumptions, a sample size of 121 cases was required for each group in order to detect a difference of 15 percentage points in hypoxemic events between both study groups with a power of 80 % using a two-sided significance rate of 5 %. Thus, a sample size of 242 patients was calculated.
!
Statistical analysis Continuous data are described by mean (standard deviation [SD]) or median (range); for categorical data absolute and relative frequencies are presented. Group comparisons of categorical data, including the primary end point of frequency of hypoxemic events, were conducted using χ² tests. If the number of expected cell counts was smaller than 5 for more than 20 % of the cells, Fisher’s exact test was used. For quantitative data, Student’s t
Intention-to-treat analysis A total of 263 patients were screened for enrolment. In 4 cases, patients failed to meet eligibility criteria; 11 patients refused to participate; and in 6 other cases written informed consent was not possible. Thus, 122 and 120 patients, respectively, were randomized to the capnography arm and standard arm. In the capnography arm 1 patient had to be excluded because endoscopic ultrasonography (EUS) was done instead of ERCP. In the standard arm 2 patients received EUS and 1 received percutaneous transhepatic biliary drainage (PTBD) instead of ERCP, which led to exclusion. Thus, 238 patients were included in the intention-to" Fig. 1). treat analysis (●
Patient, clinical and procedural characteristics: intention-to-treat Patient characteristics were well balanced with regard to age, gender, body composition and medical history with all of the ex" Table 1, P values not ploratory two-sided P values being > 0.05 (●
Klare Peter et al. Capnographic monitoring of midazolam and propofol sedation during ERCP… Endoscopy 2016; 48: 42–50
Downloaded by: University of Pittsburgh. Copyrighted material.
44
Original article
Factor
Capnography arm
Standard arm
( n = 121)
(n = 117)
Age, mean (SD), years
62.2 (14.2)
Male sex, n (%)
70 (57.9)
Body mass index, mean (SD), kg/m 2
61.7 (14.9) 57 (48.7)
25.1 (3.8)
25.5 (4.7)
Smoking (current and previous), n (%)
57 (54.8)
Alcohol abuse, current and previous, n (%)
14 (11.6)
56 (47.9) 9 (7.7)
Regular narcotic/sedative use, n (%)
18 (14.9)
14 (12.0)
Heart disease, n (%)
37 (30.6)
26 (22.2)
Lung disease, n (%)
13 (10.7)
10 (8.5)
Renal disease, n (%)
11 (9.17)
10 (8.5)
Liver disease, n (%)
38 (31.4)
35 (29.9)
Sleep apnea, n (%)
1 (0.8)
45
Table 1 Intention-to-treat analysis of monitoring with additional capnography versus standard monitoring of midazolam and propofol sedation during endoscopic retrograde cholangiopancreatography (ERCP): demographic, clinical, and procedural characteristics for patients. Data were available from all patients except where noted
2 (1.7) (Data from 116/117, 99.1 %)
I
19 (15.7)
21 (17.9)
II
53 (43.8)
53 (45.3)
III
45 (37.2)
37 (31.6)
IV
4 (3.3)
6 (5.1)
Mallampati class, n (%)
(Data from 119/121, 98.3 %)
(Data from 114/117, 97.4 %)
I
55 (46.2)
45 (39.5)
II
36 (30.3)
32 (28.1)
III
16 (13.4)
21 (18.4)
IV
12 (10.1)
16 (14.0)
Sedation problems during previous endoscopies, n (%)
3 (2.5)
5 (4.3)
Outpatient or ambulatory patients, n (%)
8 (6.6)
12 (10.3)
Benign biliary stenosis
33 (27.3)
21 (17.9)
Malignant biliary stenosis
37 (30.6)
33 (28.2)
Choledocholithiasis
22 (18.2)
31 (26.5)
Stent extraction
7 (5.8)
10 (8.5)
Biliary leaks
4 (3.3)
7 (6.0)
Indications, n (%)
Chronic pancreatitis
13 (10.7)
6 (5.1)
5 (4.1)
9 (7.7)
Cholangioscopy, n (%)
6 (5.0)
1 (0.9)
Altered anatomy, n (%)
1 (0.8)
5 (4.3)
Others
Baseline oxygen saturation, mean (SD), %
98.6 (1.5)
98.6 (1.6)
Baseline heart rate, mean (SD), beats/min
77.6 (13.7)
80.2 (16.1)
Baseline systolic blood pressure, mean (SD), mmHg
143.5 (21.8)
143.9 (21.7)
Total propofol dose, median (range), mg
400.0 (110 – 1610)
390.0 (80 – 1160) (Data from 116/117, 99.1 %)
38.0 (6 – 165)
Procedure time, median (range), min
shown). Malignant biliary stenosis was the most common indication for ERCP in both study arms. Baseline demographic, clinical, and procedural characteristics for both groups are presented in ●" Table 1.
Primary study outcome: intention-to-treat At least one episode of hypoxemia (Sao2 < 90 %) occurred in 46 patients in the capnography arm (38.0 %, 95 %CI 29.4 % – 46.9 %]) and in 52 patients in the standard arm (44.4 %, 95 %CI 35.3 % – 53.9 %) (P = 0.314 for difference in relative frequencies; OR = 1.3, 95 %CI " Table 2). In the capnography arm a total of 143 hy0.78 – 2.19; ●
38.0 (5 – 164) (Data from 116/117, 99.1 %)
poxemic events were observed compared with 158 hypoxemic events in the standard arm.
Secondary study outcomes: intention-to-treat Vital sign parameters. Apnea was detected in 78 patients in the capnography arm (64.5 %, 95 %CI 55.6 % – 73.0 %) and in 7 patients in the standard arm (6.0 %, 95 %CI 2.8 % – 11.5 %) (P < 0.001 for difference). In the capnography arm a total number of 284 events of apnea were detected. The capnography device indicated another 76 events which did not correspond to apnea clinically based upon the assessment of the independent observer. Technical problems accounted for 34
Klare Peter et al. Capnographic monitoring of midazolam and propofol sedation during ERCP… Endoscopy 2016; 48: 42–50
Downloaded by: University of Pittsburgh. Copyrighted material.
ASA class, n (%)
Original article
Table 2 Intention-to-treat analysis of prespecified outcomes in patient groups with additional capnographic monitoring or with standard monitoring of midazolam and propofol sedation during ERCP. Data were available for all patients except when noted otherwise.
Factor
Capnography arm (n = 121)
Standard arm (n = 117)
46 (38.0 %)
52 (44.4 %)
P value1
Primary end point, n (%) Oxygen saturation < 90 %
0.314
Secondary end points, n (%) Detection of apnea
78 (64.5 %)
7 (6.0 %)
Oxygen saturation ≤ 85 %
27 (22.3 %)
31 (26.5 %)
0.453
Oxygen saturation < 80 %
15 (12.4 %)
15 (12.8 %)
0.922
Oxygen supplementation increased
38 (31.4 %)
42 (35.9 %)
0.463
1 (0.8 %)
1 (0.9 %)
0.981
Assisted ventilation Death
< 0.001
0 (0.0 %)
1 (0.9 %)
0.308
Bradycardia
11 (9.1 %)
11 (8.5 %)
0.934
Hypotension
6 (5.0 %)
12 (10.3 %)
0.122
Patient cooperation rated by the endoscopist (NAS, 1 – 10), median (range)
8.0 (2 – 10) (Data available: 118/121, 97.5 %)
8.0 (1 – 10) (Data available: 113/117, 96.6 %)
0.271
Patient satisfaction (NAS, 1 – 10), median (range)
9.0 (2 – 10) (Data available: 95/121, 78.5 %)
9.5 (3 – 10) (Data available: 90/117, 76.9 %)
1
To eye-opening
5.0 (0 – 21) (Data available: 103/121, 85.1 %)
5.0 (1 – 21) (Data available: 101/117, 86.3 %)
0.841
To leaving of procedure room
7.0 (1 – 23) (Data available: 102/121, 84.3 %)
7.0 (2 – 21) (Data available: 93/117, 79.5 %)
0.689
Time from end of endoscopy, median (range) min
NAS, numerical analogue scale 1 P values derived from Student t tests, Mann-Whitney U tests, and Fisher`s exact tests as appropriate.
of these 76 false alarms (44.7 %), namely: dislocation of the nasal cannula, 11 events; blocking of the filter line, 10 events; orotracheal suctioning, 5 cases; and insertion of a nasopharyngeal tube, 8 cases. In the remaining 42 events (55.3 %) no specific cause could be identified. Severe hypoxemia (defined as Sao2 < 85 %) occurred in 27 patients (22.3 %) in the capnography arm and in 31 patients (26.5 %) in the standard arm (P = 0.453 for difference; OR = 1.26, 95 %CI 0.69 – 2.27). There were no differences between the two groups regarding the incidence rates of increased oxygen supplementation (P = 0.463), bradycardia (P = 0.934), and hypoten" Table 2). sion (P = 0.122) (● Patient cooperation, quality of sedation, and recovery. Cooperation scores as rated by the endoscopists after ERCP procedures did not differ between the study arms (8.0 [2 – 10] vs. 8.0 [1 – 10]; P = 0.271). Patient satisfaction was high in both groups and no difference could be detected between the capnography arm and the standard arm (9.0 [2 – 10] vs. 9.5 [3 – 10]; P = 1.000). The use of capnography had no influence on the time from end of endoscopy to eye-opening (5.0 min [0 – 21]) vs. 5.0 min [1 – 21]; P = 0.841) or to transfer of the patient to the recovery area (capnography 7.0 min [1 – 23] vs. standard 7.0 min [2 – 21]; P = 0.689).
guidelines. Additionally, intravenous thrombolysis was administered because of the possibility of an acute thromboembolic event. Under high catecholamine doses circulation was restored approximately 45 minutes after the beginning of cardiopulmonary resuscitation. The patient was then transferred to the intensive care unit. Unfortunately, the patient died shortly thereafter because of repeated circulatory arrest. Unblinded capnography data including end-tidal CO2 values were analyzed retrospectively. Data indicated that hypoxemia and apnea occurred simultaneously 52 minutes after the beginning of ERCP. During the following minutes of the procedure capnography showed hypoventilation. Shortly before the development of critical hypotension, bradycardia and hypoxemia end-tidal CO2 values indicated another episode of apnea. The abovementioned death was rated as a serious adverse event and was therefore reported to the Ethics Committee of the Technische Universität München, Munich, Germany. The study was stopped until written evaluation of the incident by the ethics committee was available. The committee found that the incident was not triggered by the participation in the study itself. We therefore continued inclusion in May 2011.
Per-protocol analysis Serious adverse events In April 2011, one patient in the standard arm developed severe hypoxemia approximately 52 minutes after the beginning of the ERCP procedure. Up to this point a total dose of 2.5 mg midazolam and 300 mg propofol had been given. Administration of sedatives was stopped immediately and oxygen supplementation was increased. At 6 minutes after the first event of oxygen desaturation the patient developed critical hypotension (systolic blood pressure < 90 mmHg) and bradycardia (heart rate < 50/min). Cardiopulmonary arrest then occurred. Cardiopulmonary resuscitation was started immediately and orotracheal intubation was conducted. During the following 45 minutes of resuscitation, adrenaline was given repeatedly according to current resuscitation
Of the 238 patients, who were included in the intention-to-treat analysis, 15 were excluded from the per-protocol analysis. Among these, 2 accidentally received no oxygen supplementation; 1 other patient received 4 L/min oxygen during the procedure; and 1 patient received etomidate instead of propofol. In 11 cases in the capnography arm, continuing problems using the capnography device were reported that made the performance of capnography impossible. These cases were rated as protocol violations. All 11 cases were caused by practical problems in applying capnography: continuing ongoing problems placing the sample line in 6 cases; need to exchange the oxygen probes in 3 cases; and blocked filter lines in 2 cases. Thus, 223 patients were included in the per-protocol analysis.
Klare Peter et al. Capnographic monitoring of midazolam and propofol sedation during ERCP… Endoscopy 2016; 48: 42–50
Downloaded by: University of Pittsburgh. Copyrighted material.
46
Original article
Factor
Capnography arm
Standard arm
(n = 108)
(n = 115)
Age, mean (SD), years
62.1 (14.1)
61.9 (14.8)
Male sex, n (%)
60 (55.6)
57 (49.6)
Body mass index, mean (SD), kg/m 2
24.8 (3.7)
25.5 (4.7)
Smoking (current and previous), n (%)
50 (46.3)
56 (48.7)
Alcohol abuse (current and previous), n (%)
13 (12.0)
9 (7.8)
Regular narcotic/sedative use, n (%)
18 (16.7)
14 (12.2)
Heart disease, n (%)
29 (26.9)
26 (22.6)
Lung disease, n (%)
12 (11.1)
10 (8.7)
Renal disease, n (%)
11 (10.2)
10 (8.7)
Liver disease, n (%)
34 (31.5)
34 (29.6)
Sleep apnea, n (%)
1 (0.9)
47
Table 3 Per-protocol analysis of monitoring with additional capnographic monitoring versus standard monitoring of midazolam and propofol sedation during ERCP: demographic, clinical, and procedural characteristics for patients. Data available from all patients except where noted.
1 (0.9), (Data from 114 /115, 99.1 %)
I
17 (15.7)
20 (17.4)
II
49 (45.4)
53 (46.1)
III
39 (36.1)
36 (31.3)
IV
3 (2.8)
6 (5.2)
Mallampati class, n (%)
(Data from 106 /108, 98.1 %)
(Data from 112 /115, 97.4 %)
I
52 (49.1)
45 (40.2)
II
30 (28.3)
31 (27.7)
III
13 (12.3)
21 (18.6)
IV
11 (10.4)
15 (13.4)
Sedation problems during previous endoscopies, n (%)
3 (2.8)
5 (4.3)
Outpatient or ambulatory patients, n (%)
8 (7.4)
11 (9.6)
Benign biliary stenosis
31 (28.7)
21 (18.3)
Malignant biliary stenosis
32 (29.6)
32 (27.8)
Choledocholithiasis
18 (16.7)
31 (27.0)
Stent extraction
7 (6.5)
9 (7.8)
Biliary leaks
4 (3.7)
7 (6.1)
11 (10.2)
6 (5.2)
5 (4.6)
9 (7.8)
Indications, n (%)
Chronic pancreatitis Others Cholangioscopy
6 (5.6)
1 (0.9)
Altered anatomy
1 (0.9)
5 (4.3)
Baseline oxygen saturation, mean (SD), %
98.6 (1.5)
98.6 (1.6)
Baseline heart rate mean (SD), beats/min
78.2 (13.6)
80.2 (16.2)
Baseline systolic blood pressure, mean (SD), mmHg
144.3 (20.5)
144.0 (21.7)
Total propofol dose, median (range), mg
390.0 (110 – 1610)
380.0 (80 – 1660)
37.0 (6 – 165)
Procedure time, median (range), min
38.0 (5 – 164) (Data from 114 /115, 99.1 %)
P values derived from to Student t tests, Mann-Whitney U tests and Pearson chi-squared tests as appropriate.
Patient, clinical and procedural characteristics: per-protocol Baseline demographic, clinical, and procedural characteristics for " Table 3 and showed an equitable both groups are presented in ● distribution between the groups with all of the exploratory twosided P values being > 0.05 (P values not shown).
Primary study outcome: per-protocol Hypoxemia (at least one episode of Sao2 < 90 %) was observed in 34 patients in the capnography arm (31.5 %, 95 %CI 22.9 – 41.1 %) and 51 patients in the standard arm (44.4 %, 95 %CI 35.1 – 53.5 %) " Table 4). (P = 0.048 for difference; OR = 1.73, 95 %CI 1.003 – 3) (●
Secondary study outcomes: per-protocol Vital sign parameters. Apnea was detected in 66 patients in the capnography arm (61.1 %, 95 %CI 51.6 % – 70.3 %) and in 6 in the standard arm (5.2 %, 95 %CI 2.3 % – 11 %) (P < 0.001 for difference). Severe hypoxemia (defined as Sao2 < 85 %) was observed less frequently in the capnography arm compared with the standard arm (16 [14.8 %] vs. 30 [26.1 %], P = 0.038; OR = 2.03, 95 %CI 1.03 – 3.98). The incidence of increased oxygen supplementation was lower in the capnography arm compared with controls (24 [22.2 %] vs. 41 [35.7 %], P = 0.028; OR = 1.94, 95 %CI 1.07 – 3.5]). There were no differences between incidence rates of bradycar-
Klare Peter et al. Capnographic monitoring of midazolam and propofol sedation during ERCP… Endoscopy 2016; 48: 42–50
Downloaded by: University of Pittsburgh. Copyrighted material.
ASA class, n (%)
Original article
Table 4 Per-protocol analysis of prespecified outcomes in patient groups with additional capnographic monitoring or with standard monitoring of midazolam and propofol sedation during ERCP. Data were available for all patients unless otherwise noted.
Factor
Capnography arm (n = 108)
Standard arm (n = 115)
34 (31.5 %)
51 (44.4 %)
P value1
Primary end point, n (%) Oxygen saturation < 90 %
0.048
Secondary end points, n (%) Detection of apnea
66 (61.1 %)
6 (5.2 %)
Oxygen saturation ≤ 85 %
16 (14.8 %)
30 (26.1 %)
0.038
Oxygen saturation < 80 %
7 (6.5 %)
14 (12.2 %)
0.146 0.028
Oxygen supplementation increased
< 0.001
24 (22.2 %)
41 (35.7 %)
Assisted ventilation
0 (0.0 %)
1 (0.9 %)
Death
0 (0.0 %)
1 (0.9 %)
Bradycardia
9 (8.3 %)
11 (10.2 %)
0.748
Hypotension
5 (4.6 %)
1 1
12 (10.4 %)
0.103
Patient cooperation rated by the endoscopist (NAS, 1 – 10), median (range)
8.0 (2 – 10) (Data available: 106 /108, 98.1 %)
8.0 (1 – 10) (Data available: 111 /115, 96.2 %)
0.082
Patient satisfaction (NAS, 1 – 10), median (range)
9.0 (2 – 10) (Data available: 88 /108 (81.5 %)
9.0 (3 – 10) (Data available: 89 /115 (77.4 %)
0.992
To eye-opening
5.9 (0 – 21) (Data available: 92 /108 (85.2 %)
5.6 (1 – 21) (Data available: 99 /115 (86.1 %)
0.944
To leaving of procedure room
7.8 (1 – 23) (Data available: 91 /108 (84.3 %)
7.5 (2 – 21) (Data available: 91 /115 (79.1 %)
0.944
Time from end of endoscopy, median (range) min
NAS, numerical analogue scale 1 P values derived from Student t tests, Fisher`s exact tests, and Mann-Whitney U tests, as appropriate.
dia (P = 0.748) and hypotension (P = 0.103) between the two " Table 4). groups (● Patient cooperation, quality of sedation, and recovery. Cooperation scores as rated by the endoscopist after ERCP procedures did not differ between the study groups (capnography 8.0 [2 – 10] vs. standard 8.0 [1 – 10], P = 0.082). Patient satisfaction scores were equal in the two groups (9.0 [2 – 10] vs. 9.0 [3 – 10]; P = 0.992). The use of capnography had no influence on the time from end of endoscopy to eye-opening (capnography 5.9 min [0 – 21] vs. standard 5.6 min [1 – 21], P = 0.944) or to transfer of the patient to the recovery area (capnography 7.8 min [1 – 23] vs. standard 7.5 min [2 – 21], P = 0.944).
Discussion !
The use of sedatives has improved patient comfort and acceptance with regard to endoscopy in recent years [7]. For invasive and potentially painful investigations especially, sedation is mandatory. Propofol is widely used for endoscopy but has the potential to cause cardiorespiratory side effects [8]. Capnography monitoring can detect apnea in the nonintubated patient [9], but nevertheless, the use of capnography is not established as a recognized standard for monitoring respiratory function during endoscopy [7, 10 – 13].
The influence of capnography on incidence of hypoxemic events In this study we aimed to show that capnography-directed early interventions are useful in avoiding sedation-induced hypoxemia. We found that hypoxemia incidence rates did not differ significantly between both groups in an intention-to-treat analysis. However, capnography was superior to standard oximetry in detecting apnea episodes. Capnography monitoring has been shown to help avoid desaturation in standard endoscopy procedures, such as esophagoduode-
noscopy (EGD) and colonoscopy [14, 15]. Moreover, in one trial comparing capnography versus oximetry monitoring in patients undergoing upper endosonography or ERCP, the frequency of hypoxemia was significantly reduced in the capnography arm [16] which, however, contradicts our results. However, it must be noted that in the aforementioned trial benzodiazepines and opioids were used for sedation and most of the patients (approximately 75 %) received endosonography, a less invasive investigation than ERCP [16]. One explanation for our results could be that conditions related to ERCP itself hampered effective interventions, even though good and ample attention was given to apnea events. Patients were positioned in a prone position which might act as a limiting factor when rescue maneuvers (e. g. chin-lift) are being used. We therefore assume that interventions based on the detection of apnea might have been ineffective in some cases. On the other hand we experienced some technical and practical problems which interfered with capnography measurement. Technical difficulties with capnography sampling have also been described previously [15]. In our study, we detected a remarkable number of false-positive apnea alarms given by capnography which were mainly due to technical issues such as dislocation of the nasal cannula. Difficult or invasive interventions during ERCP may lead to pain or discomfort which consequently result in agitation. This may be one out of several reasons why capnography measuring was error-prone in our trial. Stable measuring conditions for the duration of the whole procedure would be a prerequisite for the correct detection of apnea and, in consequence, for efficiently conducted countermeasures to avoid hypoxemia. These hypotheses are further supported by the results of the perprotocol analysis. When cases with persistent mismeasurement were excluded, we found a lower incidence of hypoxemia in the capnography arm compared with the standard arm. Furthermore, severe hypoxemic events also occurred less frequently in patients in the capnography arm. One possible interpretation might be that if capnography could be applied correctly without
Klare Peter et al. Capnographic monitoring of midazolam and propofol sedation during ERCP… Endoscopy 2016; 48: 42–50
Downloaded by: University of Pittsburgh. Copyrighted material.
48
Original article
Serious adverse event during the trial Unfortunately, in one patient receiving standard monitoring cardiopulmonary arrest occurred during ERCP, which subsequently led to the death of the patient. The exact cause of the event remained unknown although cardiorespiratory depression related to sedation seemed probable. The unblinded capnography data retrospectively showed that apnea episodes occurred immediately prior to the development of hypoxemia and cardiac arrest in this case. The preventability of this event remains a subject of speculation. Sedation-related death during endoscopy procedures has been shown to occur rarely [17]. Furthermore, high patient morbidity (ASA class III and higher) as well as emergency investigations are factors which predict an increased risk for severe events [18]. In our case the patient was classified as ASA II, and was suffering from biliodigestive stenosis after surgery (Whipple procedure) for carcinoma of the papilla of vater. In summary this tragic incident shows that endoscopic procedures under sedation bear a significant risk even in patients who present in a relatively good general condition.
Study limitations Our study is subject to several limitations. Neither patients nor the endoscopist and physician responsible for sedation nor the independent observer were blinded to assignment of patients. Thus there was the potential for bias regarding the assessment of apnea episodes. Secondly, as no validated scale was used to assess depth of sedation, this could have differed between the two groups and could have led to differences in the incidence of oxygen desaturation and hypoxemia. However, the mean total propofol doses did not differ between the groups, suggesting that there was no bias regarding the application of propofol. Furthermore, hypoxemic events were observed frequently without the occurrence of clinically relevant consequences. This illustrates that hypoxemia only served as a surrogate parameter. The occurrence of hypoxemia might not be interpreted as a sedationrelated adverse event in itself. One way of addressing this issue could be to primarily focus on severe hypoxemia (Sao2 < 85 %). However, it is unclear whether severe hypoxemia is more feasible in serving as an adequate surrogate for sedation-related complications. Furthermore, any fall of Sao2 < 90 % is considered to be important, as stated by gastroenterological associations [7, 11]. We therefore decided to use the latter threshold as target indicator in our trial. Finally, oxygen supplementation may have masked hypoventilation, leading to artificially high Sao2 measurements. However, current national guidelines recommend routine use of oxygen supplementation for procedural sedation during endoscopy [7].
Conclusion In summary, our study showed, that capnography can adequately detect apnea in the setting of ERCP under midazolam and propofol sedation. Based on the results of the intention-to-treat analysis, no beneficial effect of capnography-based early intervention could be demonstrated. However, capnographic measurement
was hindered in several patients because of particular circumstances during ERCP. After the exclusion of cases with technical problems, per-protocol analysis revealed a lower incidence of hypoxemia in the capnography arm compared with the standard monitoring arm. We therefore conclude that the practical application of capnography needs to be improved. If this is possible, capnography may be of value in improving patient safety in the future. In this case, the results of our study should be noted when selecting monitoring devices for routine ERCP procedures. Competing interests: Stefan von Delius, Andrea Riphaus and Till Wehrmann received material support for research purposes from Oridion Medical (Needham, Massachusetts, U.S.). Oridion Medical provided the capnography monitor (Capnostream 20) and sample lines (Smart CapnoLine Guardian), but did not participate in study design, data collection, data analysis, or manuscript preparation. No financial support was received for this study. Institutions 1 II. Medizinische Klinik und Poliklinik, Klinikum rechts der Isar der Technischen Universität München, Germany 2 Klinik für Innere Medizin I, Universitätsklinikum Ulm, Germany 3 Institut für Medizinische Biometrie, Epidemiologie und Medizinische Informatik, Universität des Saarlandes, Campus Homburg, Germany 4 Medizinische Universitätsklinik des Knappschaftskrankenhaus der Ruhr-Universität Bochum, Germany 5 Klinik für Anästhesie, Universitätsklinikum der Universität Witten/Herdecke, Helios Klinikum Wuppertal, Germany 6 Fachbereich Gastroenterologie, Deutsche Klinik für Diagnostik, Helios Klinik Wiesbaden, Germany 7 Medizinische Klinik, KRH Klinikum Agnes Karll, Laatzen, Germany
References 1 McQuaid KR, Laine L. A systematic review and meta-analysis of randomized, controlled trials of moderate sedation for routine endoscopic procedures. Gastrointest Endosc 2008; 67: 910 – 923 2 Gerstenberger PD. Capnography and patient safety for endoscopy. Clin Gastroenterol Hepatol 2010; 8: 423 – 425 3 Vargo JJ, Zuccaro GJr, Dumot JA et al. Automated graphic assessment of respiratory activity is superior to pulse oximetry and visual assessment for the detection of early respiratory depression during therapeutic upper endoscopy. Gastrointest Endosc 2002; 55: 826 – 831 4 Burton JH, Harrah JD, Germann CA et al. Does end-tidal carbon dioxide monitoring detect respiratory events prior to current sedation monitoring practices? Acad Emerg Med 2006; 13: 500 – 504 5 Deitch K, Miner J, Chudnofsky CR et al. Does end tidal CO2 monitoring during emergency department procedural sedation and analgesia with propofol decrease the incidence of hypoxic events? A randomized, controlled trial Ann Emerg Med 2010; 55: 258 – 264 6 R Core Team. R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing; 2014: http://www.Rproject.org/ 7 Riphaus A, Wehrmann T, Weber B et al. [S3-guidelines – sedation in gastrointestinal endoscopy]. Z Gastroenterol 2008; 46: 1298 – 1330 DOI 10.1055/s-2008-1027850 8 Bell GD. Preparation, premedication, and surveillance. Endoscopy 2004; 36: 23 – 31 9 Manifold CA, Davids N, Villers LC et al. Capnography for the nonintubated patient in the emergency setting. J Emerg Med 2013; 45: 626 – 632 10 Lichtenstein DR, Jagannath S, Baron TH et al. Sedation and anesthesia in GI endoscopy. Gastrointest Endosc 2008; 68: 815 – 826 11 Dumonceau JM, Riphaus A, Aparicio JR et al. European Society of Gastrointestinal Endoscopy, European Society of Gastroenterology and Endoscopy Nurses and Associates, and the European Society of Anaesthesiology Guideline: Non-anesthesiologist administration of propofol for GI endoscopy. Endoscopy 2010; 42: 960 – 974 12 ASGE Technology Committee Gottlieb KT, Banerjee S, Barth BA et al. Monitoring equipment for endoscopy. Gastrointest Endosc 2013; 77: 175 – 180
Klare Peter et al. Capnographic monitoring of midazolam and propofol sedation during ERCP… Endoscopy 2016; 48: 42–50
Downloaded by: University of Pittsburgh. Copyrighted material.
any disruption, this can avoid hypoxemia in patients receiving ERCP under propofol and midazolam sedation. We therefore conclude that efforts should be directed to improve the application of capnography in ERCP settings. In particular, there is a need to improve sampling under difficult conditions. If this were possible, further studies would be welcome to verify our assumption.
49
Original article
13 American Society for Gastrointestinal Endoscopy, American Gastroenterological Association, American College of Gastroenterology. MultiSociety Statement: Universal adoption of capnography for moderate sedation in adults undergoing upper endoscopy and colonoscopy has not been shown to improve patient safety or clinical outcomes and significantly increases costs for moderate sedation. http://www.asge.org/ WorkArea/showcontent.aspx?id=14226 [Accessed: January 10, 2015] 14 Lightdale JR, Goldmann DA, Feldman HA et al. Microstream capnography improves patient monitoring during moderate sedation: a randomized, controlled trial. Pediatrics 2006; 117: e1170 – e1178 15 Beitz A, Riphaus A, Meining A et al. Capnographic monitoring reduces the incidence of arterial oxygen desaturation and hypoxemia during propofol sedation for colonoscopy: a randomized, controlled study (ColoCap Study). Am J Gastroenterol 2012; 107: 1205 – 1212
16 Qadeer MA, Vargo JJ, Dumot JA et al. Capnographic monitoring of respiratory activity improves safety of sedation for endoscopic cholangiopancreatography and ultrasonography. Gastroenterology 2009; 136: 1568 – 1576 17 Rex DK, Deenadayalu VP, Eid E et al. Endoscopist-directed administration of propofol: a worldwide safety experience. Gastroenterology 2009; 137: 1229 – 1237 18 Behrens A, Labenz J, Schuler A et al. [How safe is sedation in gastrointestinal endoscopy? A multicentre analysis of 388,404 endoscopies and analysis of data from prospective registries of complications managed by members of the Working Group of Leading Hospital Gastroenterologists (ALGK)] Z Gastroenterol 2013; 51: 432 – 436
Downloaded by: University of Pittsburgh. Copyrighted material.
50
Klare Peter et al. Capnographic monitoring of midazolam and propofol sedation during ERCP… Endoscopy 2016; 48: 42–50
