Neuroradiological procedures have advanced greatly in the last 10 yr. Our neuroradiology department undertakes investigative procedures for a neurosurgical centre. Digital subtraction cerebral angiography has now universally replaced the old film cassette system and results in a considerably shorter angiography time. However, in our institution patients are routinely given anaesthetist-administered sedation during the procedure, for radiological convenience and patient comfort. The benefits of anaesthetist-administered sedation have been recognized, admittedly mainly for intervention, rather than straightforward diagnostic procedures .
Previous work in our department in 1989  showed that a sedation regimen using a propofol infusion combined with intermittently administered boluses of fentanyl gave significantly earlier first recall than a regimen using boluses of fentanyl and midazolam. The objective of the present study was to reassess, 10 yr after our first study, the effectiveness of sedation for neuroradiological digital subtraction cerebral angiography, using three techniques currently in use in the department. These were an infusion of propofol mixed with alfentanil, propofol infusion with intermittent fentanyl and boluses of fentanyl and midazolam. The latter two techniques were those used in our earlier study .
Following Local Research Ethics Committee approval, we recruited 120 adult patients aged between 18 and 75 yr. Informed written consent was obtained from all patients who were all scheduled for cerebral angiography, either for suspected vascular aneurysm, arteriovenous malformation or intracranial tumour. They were graded ASA I-III. We excluded any patients who had pre-angiography confusion, disorientation, dysphasia or aphasia.
Patients were given a premedication of lorazepam 1-2 mg orally in the ward, or midazolam 1-2 mg was given intravenously (i.v.), 5 min before the start of the procedure. They were then randomized to one of three sedation techniques by sealed envelopes. All patients were sedated by one of the authors.
Group PA patients were sedated with propofol and alfentanil. Propofol 400 mg and alfentanil 2 mg were added to the same 50 mL syringe and delivered by a syringe pump. The infusion was initially commenced at 99 mL h−1 for 2-5 min and then reduced to 10-30 mL h−1 to achieve an appropriate level of sedation to be discussed later. The correct time to reduce the rate was when the patient's eyelids began to droop, not the desired sedation end-point, because of the time for administered drug to reach the effector site.
Group FP patients were sedated with an initial bolus of fentanyl 25-50 μg and a propofol 1% infusion. This again was initially commenced at 99 mL h−1 and then reduced to 10-30 mL h−1 to the appropriate level of sedation. Further increments of fentanyl 25-50 μg were given as required. This need for additional analgesia was indicated by the patient spontaneously moving, but still appearing otherwise adequately sedated.
Group FM patients were sedated with incremental doses of fentanyl and midazolam; the initial dose of fentanyl was 25-75 μg and midazolam 1-2 mg. These were repeated as necessary to gain the level of sedation required. Increments were often given together, but an apparently adequately sedated patient who was still spontaneously moving indicated the need for more analgesia.
We aimed for a sedation end-point, which gave a calm, still patient in a situation of one-way communication, e.g. obeying commands. Most patients responded with spontaneous eye opening or some movement in response to skin preparation and infiltration of the groin with local anaesthetic. Femoral artery cannulation for the angiogram catheter was then performed. Ideally, patients remained calm and still thereafter, but usually able to respond, e.g. if the head needed repositioning. However, formal clinical assessment of the sedation level throughout the procedure was impractical, as the assessment itself would disturb an otherwise peaceful patient.
Supplementary oxygen was given to all patients, via nasal cannulae placed in the mouth, rather than the nose, because this tubing does not give such a marked outline on the films compared to an MC-type facemask; positioning the prongs between the lips is less irritant than placement in the nose. All patients were monitored using non-invasive blood pressure, electrocardiography and pulse oximetry throughout the procedure.
At the end of the procedure, administration of sedation was stopped and the time to orientation (giving correct date of birth and day of the week) was noted. Patients remained on the angiogram table while pressure was applied to the groin where the arterial puncture had been made. Following this, patients were transferred back to the trolley prior to returning to the ward, without any further need for a formal recovery period.
Each patient was interviewed by one of the authors on the morning following the procedure. They were asked their last memory of events leading up to the angiogram, what their first memory was after the conclusion of the angiogram and if they had any awareness of the procedure itself. Few patients were aware of the actual time of first recall, so we asked what was going on and where the patient was. This was to determine if first recall was while still in the radiography department or not until back in the ward. The person conducting the interview was often not blind to the group to which the patient belonged. However, the interview was conducted on a semi-structured basis, with information obtained on first positive response to a sequence of questions.
We excluded from analysis any patient we were unable to interview after the angiogram. This included patients who had subsequently gone to theatre before the follow-up interview could take place. In addition, any patient who had a post-angiogram cerebral event, such as a re-bleed, was excluded and some patients could not be followed up because of administrative reasons. Data was analysed using ANOVA or Kruskall-Wallis non-parametric ANOVA as appropriate.
No procedure had to be abandoned or converted to a general anaesthetic because of unsatisfactory sedation preventing the angiogram being completed. However, of the 120 patients recruited, 32 (28%) were excluded from the analysis because of failure of follow-up interview. Four had already gone to the operating room for definitive surgery, six had another bleed after the angiogram and 22 could not be followed up because of unavailability of either author the following day. Thus, 88 patients completed the study including the follow-up interview and are included in the analysis of results. The ages of the patients were similar between the three groups, as was the duration of the angiography procedure (Table 1). The mean doses of the sedation drugs used are also shown in Table 1.
The recovery data are shown in Table 2. There was no statistically significant difference in time to eye opening or orientation in all three groups, although the PA group showed a trend to more rapid recovery. There was no correlation between the duration of the procedure and time to orientation, within each sedation group, or when all patients were grouped together. The numbers of patients whose first recall, after the procedure, of still being in the neuroradiology department were similar between groups. All patients were able to recall events prior to the sedation.
Altogether, there were 18 patients (20%) out of the 88 who had some awareness of the angiogram procedure itself. This troubled none of these patients. Although 50% (nine) of these patients were from the FP group, there was no statistical significance in the distribution of these patients between the three groups.
We have shown that successful sedation can be achieved for neuroradiological procedures, with satisfactory conditions during the angiogram and an acceptable time to return of full orientation by any of three different anaesthetist-administered sedation regimens. The distinction between sedation and anaesthesia in the radiological environment, with regard to administration by those other than anaesthetists has been clearly stated . The need for sedation during neuroradiological procedures has been demonstrated [1,4], although one article suggested that adverse events were just as frequent when sedation was administered by a radiologist, rather than anaesthetist . However, numbers in this article were far too small to justify the authors' claim of acceptable safety in the hands of a single practitioner radiologist. The desirability of a person to administer the sedation and monitor the patient, who is additional to the practitioner performing the procedure, has been clearly stated in ophthalmic surgery [6,7].
Neuroradiological investigations have changed dramatically over the last 10 yr. In particular, digital subtraction cerebral angiography allows visualization of the vessels in real time as the contrast medium is injected, as well as instant replay on demand, thus saving the wait to have conventional films developed and processed. We found procedures were on average 12.5 min shorter than in our previous sedation study (37 min vs. 48.5 min) which we attribute entirely to the use of digital subtraction cerebral angiography radiographic equipment. Despite this, the times to orientation of the three groups (3.7, 4.3 and 5.1 min) are comparable to the times to orientation in our previous study (6.2 and 4.2 min). This is in spite of the shorter time for the procedure. Furthermore, there was no significant correlation between duration of procedure and time to orientation, either within groups, or for the whole patient group combined. We feel that the duration of the procedure has little influence on the time to orientation.
We also previously found the midazolam-fentanyl incremental technique superior to a propofol infusion technique for initial wakeup (time to orientation), but inferior with regard to early first recall (recall while still in the radiography department). However, in our present study, there is no marked difference between any of the three groups in time to orientation, or numbers with first recall in the radiography department. All three sedation techniques appear to give comparable speed of recovery.
It is important and desirable to have rapid orientation, appropriate behaviour and wakefulness after cerebral angiography, if for no other reason than the ability to identify any neurological deterioration in the patient after the angiogram; thus, general anaesthesia is now rarely used. However, patients are often frightened and do not want to be aware or have any recall of the procedure. It can also be easier for the radiologist, if the patient does not continually move or fidget. Thus, a fine balance must be achieved between adequate sedation and oversedating a patient. Our study has shown no clinically significant differences between the three techniques that we routinely use. One study comparing midazolam and propofol infusions for radiological sedation  found no difference between the two techniques. Both groups were given analgesia as required (fentanyl) but a high incidence of pain, inappropriate movements and apparent oversedation were seen. Total drug utilization was higher than our study, but the procedure duration was much longer.
In our previous study, the propofol-fentanyl group were significantly more likely to have their first memory in the radiography department. In addition, only one patient in that study had any awareness of the procedure. In the current study, similar proportion of patients in each group had early recall in the radiography department, but many more patients than in our previous study overall had some awareness of the procedure itself, possibly suggesting that we are now using a lighter level of sedation. We made a clinical judgement regarding the sedation depth in both this and the previous study, the most important components being a safe patient who remained calm and still during the procedure. More formal clinical assessment of sedation, such as demonstrating the ability to obey commands, would have ruined the tranquillity of the patient's sedation and disrupted the angiogram and the use of physiological methods, such as electroencephalography, are neither validated nor possible during cerebral angiography, as the electrodes could obscure part of the cerebral vasculature.
We ensured that all patients received a modest dose of anxiolytic agent, either in the form of an oral lorazepam premedication, or a small dose of midazolam i.v. just before the procedure. All patients could remember arriving in the department and events, such as placement of the venous cannula. Thus, the anxiolytic agent has probably not contributed greatly to any delay in recovery. However, we have found that anxiolysis is an essential component of successful sedation. Without it, a patient can be apparently calm and still, but without any external stimulus and apparently at random, moves and tries to communicate every few minutes. This cannot be satisfactorily overcome by merely increasing the other sedation drugs. Vinik and colleagues  found a profound hypnotic synergism between propofol, midazolam and alfentanil, as well as with midazolam plus alfentanil when used for induction of anaesthesia. However, the lack of significant additional benefit from the analgesic component found by Vinik's group was because they were only assessing an end-point of unconsciousness. Our sedation technique needs to provide analgesia because patients are expected to remain still for a period of time, lying on a firm angiogram table, and recover quickly. Vinik and colleagues were not concerned with speed of recovery.
The need for further increments of analgesia in our patients could be distinguished from the need for additional hypnotic or anxiolysis. Whereas the need for anxiolysis was spontaneous and apparently random attempts of the patient to communicate, the indication for analgesia was by movements similar to fidgeting, due to discomfort induced by lying on the firm surface of the angiogram table. This fidgeting was not related to events during the angiogram. Thus, it was usually possible to distinguish between the three components of sedation needed by the patient at any one time. The ability to distinguish the need for further doses of each sedation component was only acquired by experience with the sedation techniques.
In conclusion, we have found that satisfactory anaesthetist-administered sedation can be provided for cerebral angiography by either infusion or incremental techniques. However, we feel that the success and safety of a sedation technique does depend on the skill and experience of the administrator. These sedation techniques are only suitable for safe use by an anaesthetist.
We thank Dr P. W. Duncan who facilitated Dr Bewlay's post from the UK National Health Service funds.
1. Luginbuhl M, Remonda L. Interventional neuroradiology. Recent developments and anaesthesiologic aspects. Minerva Anesthesiol
2. Allen MWB, Laurence AS, Gunawardena WJ. A comparison of two sedation techniques of neuroradiology. Eur J Anaesthesiol
3. Sedation and Anaesthesia in Radiology; Report of a Joint Working Party.
London, UK: Royal College of Anaesthetists and Royal College of Radiologists, 1992.
4. Lind LJ, Mushlin PS. Sedation, analgesia and anesthesia for radiologic procedures. Cardiovasc Intervent Radiol
5. Manninen PH, Gignac EM, Gelb AW, Lownie SP. Anesthesia for intervention radiology. J Clin Anesth
6. Report of the Joint Working Party on Anaesthesia in Ophthalmic Surgery.
London, UK: Royal College of Anaesthetists, College of Ophthalmologists, 1993.
7. Recommendations for Standards of Monitoring During Anaesthesia.
London, UK: Association of Anaesthetists of Great Britain and Ireland, 1998.
8. Manninen PH, Chan ASH, Papworth D. Conscious sedation for interventional neuroradiology: a comparison of midazolam and propofol infusion. Can J Anaesth
9. Vinik HR, Bradley EL, Kissin I. Triple anesthetic combination: propofol-midazolam-alfentanil. Anesth Analg