Colonoscopy is a short-lasting but sometimes painful procedure and a wide variety of medications have been used for sedation during it . When anaesthetists provide care for colonoscopy patients, propofol is usually the drug of choice because its use is associated with a rapid recovery profile and low incidence of side-effects . Although propofol is an excellent hypnotic, it is not an analgesic. Consequently, high doses are often required to enable optimal conditions for the procedure, sometimes resulting in cardiovascular and respiratory depression and loss of protective reflexes.
Controversy exists about the need for sedation and amnesia during colonoscopy . Our hypothesis was that if we succeeded in relieving pain adequately during colonoscopy, sedation would no longer be required. The use of remifentanil in this setting has advantages over other opioids because of its rapid onset and offset times, allowing rapid titration to the individual patient's requirements and to the intermittent pain of colonoscopy.
This randomized, open, prospective study was designed to test the hypothesis that a continuous infusion of remifentanil during colonoscopy can provide patient comfort, safety and recovery characteristics similar to those obtained with a propofol infusion.
The study was approved by the hospital Ethics Committee, and written informed consent was obtained from all patients. Forty 18–65-yr-old outpatients scheduled for complete colonoscopy entered the study. Patients with pulmonary, cardiovascular, hepatic, renal, haematological, neurological or metabolic diseases were excluded. Also excluded were patients with a history of allergic reactions to any of the study drugs, patients chronically receiving opioid analgesics or sedative medication, and patients who had undergone general anaesthesia within 7 days before the study.
Patients received no preanaesthetic medication. Preoperatively, they were asked to complete two baseline psychometric tests. To assess cognitive function, a Digit Symbol Substitution Test (DSST) , asking them to match numbers and symbols over 60s, was used. Psychomotor function was evaluated with a Trieger Dot Test (TDT) . The patients had to connect a series of dots and the test was scored by adding the number of dots not touched by the connecting pencil line to the time (s) it took to complete the drawing. Patients were then randomly assigned to one of two groups. Assignment to the groups was determined by random drawing of consecutively numbered envelopes containing the labels ‘remifentanil’ or ‘propofol’. In the remifentanil group (Group R), a bolus dose of remifentanil 0.5 μg kg−1 was given over 30 s, followed by a continuous infusion (reconstituted to a concentration of 10 μg mL−1) at 0.2 μg kg−1 min−1. Supplemental doses of 0.25 μg kg−1 were given if the patient experienced pain. When the colonoscope had reached the caecum and the physician had started to withdraw the instrument, the infusion rate was decreased to 0.1 μg kg−1 min−1. In the propofol group (Group P), anaesthesia was induced with propofol 1 mg kg−1 and maintained with an infusion rate of 10 mg kg−1h−1. When lightening of anaesthesia was observed (e.g. voluntary movements), an additional dose of propofol 0.5 mg kg−1 was given. When the endoscope had passed the hepatic flexure of the colon, the maintenance dose was decreased to 5 mg kg−1h−1. During the withdrawal phase of the colonoscopy, the infusion rate was decreased to 2.5 mg kg−1h−1. In both groups, the maintenance infusion rate was decreased by 50% if a ‘safety end-point’ occurred. Safety end-points included a respiratory rate (RR) <6 breaths min−1 for at least 1 min, expired carbon dioxide (ETCO2) >7.3 kPa, SPO2 < 95%, heart rate (HR) <40 beats min−1 or mean arterial pressure (MAP) <60mmHg. Infusions were discontinued at the completion of colonoscopy. In all cases, the same professional staff (A. M. and L. H.) performed anaesthesia.
During the examination, standard monitoring included electrocardiography, pulse oximetry, capnography and non-invasive measurement of blood pressure (Cardiocap®; Datex, Helsinki, Finland). Patients were allowed to breathe spontaneously. Supplemental oxygen (5 L min−1) was supplied by a facemask, modified to permit monitoring of ETCO2, RR and the respiratory pattern.
Haemodynamic and respiratory variables, including non-invasive arterial pressure, heart and RRs, ETCO2 and SPO2, were recorded before administration of the anaesthetic agent (baseline) and at 1 min intervals for 10 min after induction. Subsequent measurements were performed at 2.5 min intervals until the end of anaesthesia. The number of supplemental doses, the number of changes in the infusion rate and the total dose of remifentanil or propofol required during the procedure were recorded. Hypoventilation (RR <8 breaths min−1 for at least 1 min), apnoea (the absence of ventilatory effort for at least 30 s), airway obstruction, hypotension (a decrease of MAP by >30% from awake values), hypertension (an increase of MAP by >30% from awake values), dysrhythmias and any other adverse events or experiences were recorded.
Immediate recovery was assessed using the Steward Post Recovery Score (SPRS) . The times at which the patients spontaneously opened their eyes, were able to follow the command to squeeze the investigator's hand and were able to state their date of birth were recorded. At 5, 15 and 30 min after anaesthesia, the DSST and TDT were repeated. All tests were performed with the patients sitting upright in bed. The order of each test was held constant (DSST, TDT) and was the same for all patients.
Thirty minutes after the end of anaesthesia, patients were interviewed for amnesia and side-effects. They provided a subjective assessment of their general feeling of ‘well-being’ by scoring on a 100 mm visual analogue scale (VAS). Discharge criteria required that patients were awake and alert with stable vital signs, had no intractable side-effects and could walk without assistance. However, all patients were asked to stay for at least 30 min.
The required sample size was derived from a power analysis to demonstrate a difference in time to complete recovery (SPRS 6) of 3 min with an estimated SD = 2.5 min, α = 0.05 two-tailed and a power = 80%. Statistical analysis was performed using SPSS® statistical software (SPSS Inc., Chicago, IL, USA). The U-test was used to compare the patient characteristics data, VAS scores and the immediate recovery data. Changes over time were evaluated with twofactor ANOVA for repeated measures including a Bonferroni post hoc test. P < 0.05 was considered as significant. The data are the mean (SD), median (interquartile range (range)) or number (proportion), as appropriate.
The two study groups had similar patient characteristics (Table 1). Table 2 shows the dosage requirements for propofol and remifentanil, and the respective adaptations of their doses.
In Group R, RR was lower and ETCO2 higher compared with baseline and with Group P. In the Group P, RR was significantly higher than baseline and ETCO2 remained constant (Fig. 1). There was no statistical difference in SPO2 values between the two groups; SPO2 was >95% at all times.
Mean systolic arterial pressures (SAP), diastolic arterial pressures (DAP) and MAP for Group P were significantly lower than for Group R. In Group R, arterial pressures were not different from baseline. In Group P, SAP and MAP were significantly lower than baseline after 2 min of propofol infusion until the end of the procedure. Mean heart rate in Group P was significantly lower than in Group R, although in both groups the rate was not significantly different from baseline.
During remifentanil administration, all patients were mildly sedated, gave a lethargic response to verbal commands given in a normal tone of voice and had mild ptosis or a ‘glazed’ appearance of the eyes. In Group P, all patients were unconscious and verbal contact was lost.
In Group P, the following emergence times were recorded after discontinuation of the infusion: time to opening of the eyes, 2 min (1–3.2 (1–7)), following commands 2 min (1–3.2 (1–7)) and recalling the date of birth 2 min (1–4 (1–10)). The maximum Steward Post Recovery Score of 6 was reached in Group P after a median duration of 3 min (1.7–4 (1–10)). In Group R, all patients remained responsive throughout the procedure. Every patient in Group R had an SPRS of 6 at the moment of discontinuation of the infusion.
Performance on the DSST is shown in Figure 2. Group R performed better than Group P at all time points. The scores in Group P reached preanaesthetic values after 15 min. In Group R, patients had significantly better results compared with baseline at 15 and 30 min. Patients in Group R showed no disturbances in the TDT (Fig. 3). The TDT score was significantly worse in Group P at 5 and 15 min, in reference both to baseline and to Group R.
Adverse events for each treatment group are shown in Table 3. The most common adverse events were respiratory side-effects, which occurred in 55% of patients in Group R and in 45% of patients in Group P. One patient in Group R experienced severe abdominal pain during the procedure and could not be managed with remifentanil alone. She received a bolus of propofol (2 mg kg−1) and the procedure was continued without further problems. From this moment on, this patient was withdrawn from the study, and any further efficacy assessments for that patient were treated as missing data. In Group R, one 42-yr-old patient with no history of cardiovascular disease developed bradycardia with ventricular bigeminy. After atropine 0.5 mg intravenously, the patient recovered without sequelae.
In the interview before discharge, all patients stated they were satisfied with the anaesthetic they received. The VAS score was significantly higher for Group P than for Group R (96 ± 7 versus 77 ± 21, respectively; P < 0.001).
Thirty minutes after the end of anaesthesia, all patients, except two in each group, were judged to be ‘fit for discharge’. In Group R, one patient was not discharged because of vomiting and heaviness in the head and the other because of nausea and sweating. In Group P, one patient was very emotional and therefore not discharged, and one still felt too drowsy.
This study demonstrates that a continuous infusion of remifentanil as the sole agent is a useful and relatively safe technique for carefully monitored anaesthesia care in patients undergoing colonoscopy when administered by trained professionals. Very few studies have reported on the use of remifentanil as the sole agent during surgical or diagnostic procedures. Infusion of remifentanil is considered a satisfactory technique for oocyte retrieval and during placement of ophthalmological nerve blocks [7–9]. One study reports on the use of remifentanil versus meperidine, both combined with midazolam, during colonoscopy . The authors concluded that both opioids were equally well tolerated and suggested that the higher verbal pain and anxiety scores in the remifentanil group were caused by the fact that the remifentanil and/or midazolam doses used (228 ± 243 μg and 2.9 ± 1.2 mg, respectively) may have been insufficient in some patients. The authors did not evaluate recovery characteristics and cognitive or psychomotor function in their patients.
We decided to compare remifentanil with propofol because propofol-based anaesthesia is commonly used in our department during colonoscopy. Several studies have compared remifentanil and propofol infusions, but only as adjuncts to local or regional anaesthesia [9,11–13]. In these studies, remifentanil was used to provide sedation. Sedation is not the primary action of opioids and could reasonably be described as a sideeffect. It is therefore not surprising that in this setting remifentanil, even at low doses, is associated with a high degree of respiratory depression.
The high dose of remifentanil used in our study was based on earlier investigations [7,14] and on our own experience of patients undergoing colonoscopy. We started at an infusion rate of 0.2 μg kg−1 min−1 If pain relief was inadequate, supplemental doses were used since these will rapidly increase plasma and effect-site concentrations, whereas an increase in infusion rate will require several minutes to achieve a new steady-state concentration. The dose regimen of propofol was based on simulated plasma concentrations of propofol. Our initial infusion profile revealed a predicted plasma concentration between 3.5 and 4 μg mL−1, being around the Cp50 for loss of consciousness, as described by Smith and colleagues . As in the remifentanil group, supplemental doses were used in the propofol group to increase rapidly the plasma concentrations of propofol when clinically indicated.
Despite the high infusion rates of remifentanil in the present study, the incidence of respiratory depression is comparable with earlier studies where lower doses of remifentanil were used [11–13]. Smith and colleagues  reported decreases in the remifentanil infusion rate (mean 0.095 ± 0.020 μg kg−1 min−1) because of a low RR (<8 breaths min−1) in 41% of their patients. In our study, this incidence rate was 25%. This can be attributed to the fact that in our study no concomitant medication was used and to the stimulation caused by the procedure itself. Babenco and colleagues  assessed the time-course of the ventilatory depression after a bolus dose of remifentanil. After 0.5 μg kg−1 remifentanil given over 5 s, the peak effect was observed after 2 min and recovery to 60% of baseline was reached within 5 min of the bolus dose. In the present study, we administered bolus doses whenever the patient experienced pain. Waiting at least 2 min before giving a new dose would have resulted in a lower incidence of respiratory depression, but the aim here was to maximize analgesia and enable optimal conditions for the procedure. To achieve this objective, we sometimes needed frequent supplemental doses of remifentanil. The respiratory events were resolved simply by asking the patient to breathe deeply. In five patients, the maintenance rate had to be decreased by 50% because it was very difficult to maintain reliable spontaneous ventilation. In the dose given, propofol also had a high incidence of respiratory side-effects. Although there was no need to decrease the infusion rate, nine patients (45%) needed temporary airway support (jaw thrust).
The haemodynamic responses to both propofol and remifentanil are identical to those in other studies . Although arterial pressure and heart rate decreased significantly during the procedure in the propofol group, this was never considered clinically relevant.
It is known that very little discomfort is experienced during the withdrawal phase of the colonoscopy. Therefore, remifentanil and propofol were titrated downward near the completion of the procedure and this approach resulted in shorter emergence times compared with other studies .
Psychomotor testing was included in our protocol because it provides a more sensitive index of patient recovery than clinical assessment alone . Our results show that cognitive and psychomotor impairment was greater and more prolonged in the propofol group. In the remifentanil group, patients had significantly better results for the DSST at 15 and 30 min compared with baseline. A possible explanation for this unexpected result is a training effect, despite efforts to eliminate this by using different symbols for each test. Alternatively, patients might be more relaxed after the procedure and more able to concentrate on the test, as suggested spontaneously by two of the patients. Our study can be criticized in that patients in both groups were not at the same anaesthetic depth during the procedure. However, our study was designed to mimic a clinical situation in which the drugs are titrated to patient comfort and procedure tolerance.
Patient satisfaction was significantly higher in the propofol group. Some patients in the remifentanil group reported apprehension about remaining fully conscious and aware of events during their examination. In addition, the higher incidence of postoperative side-effects may contribute to the lower VAS score in the remifentanil group. However, it is probable that the euphoric-like state induced by propofol accounts for the observed difference .
In conclusion, remifentanil can be used as the sole agent for carefully monitored anaesthesia care during colonoscopy when administered by trained professionals. However, anxiety, apprehension and fear can greatly enhance the severity of pain or discomfort experienced. Consideration should be given to reserving this opioid for well-motivated patients who accept the fact that they will be conscious and aware of events during their examination. In these well-motivated patients, the establishment of a safe and effective patient-controlled delivery system for remifentanil needs to be investigated to allow those patients undergoing colonoscopy to adjust their dose on an ‘as needed’ basis.
1. Jamieson J. Anesthesia and sedation in the endoscopy suite? (influences and options). Curr Opin Anaesthesiol
2. Meisel M. Utilisation du Diprivan®
pour les endoscopies digestives. Ann Fr Anesth Réanim
3. Hutchinson RC, Kenny GNC. Sedation for endoscopy. Curr Opin Anaesthesiol
4. Lezak MD. Neuropsychological Assessment.
New York, USA: Oxford University Press, 1983.
5. Newman MG, Trieger N, Miller JC. Measuring recovery from anaesthesia - a simple test. Anesth Analg
6. Steward DJ. A simplified scoring system for the postoperative recovery room. Can Anaesth Soc J
7. Wilhelm W, Biedler A, Hammadeh ME, Fleser R, Grüneβ V. Infusionsanalgesie mit Remifentanil. Anaesthesist
8. Ahmad S, Leavell ME, Fragen RJ, Jenkins W, Roland CL. Remifentanil versus alfentanil as analgesic adjuncts during placement of ophthalmologic nerve blocks. Reg Anesth Pain Med
9. Holas A, Krafft P, Marcovic M, Quehenberger F. Remifentanil, propofol
or both for conscious sedation during eye surgery under regional anaesthesia. Eur J Anaesthesiol
10. Greilich PE, Virella CD, Rich JM, et al.
Remifentanil versus meperidine for monitored anesthesia care: a comparison study in older patients undergoing ambulatory colonoscopy
. Anesth Analg
11. Smith I, Avramov MN, White PF. A comparison of propofol
and remifentanil during monitored anesthesia care. J Clin Anesth
12. Mingus ML, Monk TG, Gold MI, Jenkins W, Roland C and the Remifentanil 3010 Study Group. Remifentanil versus propofol
as adjuncts to regional anesthesia. J Clin Anesth
13. Lauwers MH, Vanlersberghe C, Camu F. Comparison of remifentanil and propofol
infusions for sedation during regional anesthesia. Reg Anesth Pain Med
14. Litman RS. Conscious sedation with remifentanil during painful medical procedures. J Pain Symptom Manage
15. Smith C, McEwan AI, Jhaveri R, et al.
The interaction of fentanyl on the Cp50
for loss of consciousness and skin incision. Anesthesiology
16. Babenco HD, Conard PF, Gross JB. The pharmacodynamic effect of a remifentanil bolus on ventilatory control. Anesthesiology
17. Dubois A, Balatoni E, Peeters JP, Baudoux M. Use of propofol
for sedation during gastrointestinal endoscopies. Anaesthesia
18. Black ML, Hill JL, Zacny JP. Behavioral and physiological effects of remifentanil and alfentanil in healthy volunteers. Anesthesiology
19. Zacny JP, Lichtor JL, Coalson DW, et al.
Subjective and psychomotor effects of subanesthetic doses of propofol
in healthy volunteers. Anesthesiology