Closed Loop Control of Intravenous Agents
Gavin NC Kenny, BSc(Hons),
MB ChB, MD, FRCA, FANZCA
Professor and Head of
Department
Glasgow University
Department of Anaesthesia
We all practice anaesthesia but as yet we have no specific method by
which the level of anaesthesia can be assessed. Some would maintain that
anaesthesia is an all-or-none phenomenon and that there is no gradation. Others
would support the view that, while the loss of consciousness may well be a
single transition, different levels of the subsequent state of anaesthesia
allow different levels of responsiveness following stimulation of the patient.
A variety of different techniques have been proposed to provide an
objective assessment of anaesthetic depth. These range from the most simple
clinical observations to complex analyses of the electrical activity of the
brain. The bispectral index is one such technique which has been developed and
subjected to several clinical trials. Another technique is the auditory evoked
response which measures the response of the brain to auditory stimuli. The
principal difficulty in assessing any technique to measure anaesthetic depth is
that there is no accepted gold standard available to define the state of
anaesthesia. If we define the state of consciousness as responding to command,
then it is recognised that lack of recall for events during anaesthesia does
not guarantee that the patient was unconscious at the time. Patients have been
reported to respond to a command but to have no recollection of this
afterwards.1 This state has been called
‘general amnesia’ and illustrates the difficulty of detecting when patients
were truly anaesthetised.
If we define anaesthesia as a loss of response to command then we can
assess any monitoring system for anaesthesia by taking subjects from
consciousness to loss of consciousness and back. Changes in the signal during
these transitions must be sufficiently large that we have confidence in
detecting a true change in the clinical state of our patients from awake to
asleep and vice versa. It must function satisfactorily in all patients
and the system must be sufficiently robust to operate reliably in the operating
theatre environment. Clearly, this approach can only be used when the subjects
are either not undergoing surgery or else have a completely satisfactory local
block in place. We studied 12 patients undergoing orthopaedic surgery with a
local block.2; 3. Auditory evoked potentials
were monitored and patients were repeatedly allowed to awaken during surgery.
The auditory evoked potentials provided a clear indication of the state of
consciousness. A further study compared the AEPIndex with the
spectral edge, median frequency and bispectral index during repeated
transitions from consciousness to unconsciousness. Of the four measurements,
only the AEPIndex demonstrated a significant difference between all
mean values one minute before recovery of consciousness and all mean values one
minute after recovery of consciousness. Our findings suggest that of the four
electrophysiologic variables, the AEPIndex is best at distinguishing
the transition from unconsciousness to consciousness.4
The assessment of a monitor for anaesthetic depth during surgery
remains an outstanding problem. Anaesthetists would agree generally that
patient movement during surgery represents an inadequate level of anaesthesia.
However, there is no report of any patient having experienced awareness in MAC
studies which are designed deliberately to have 50% of patients move. One
possible gold standard is that any monitor of anaesthetic depth should allow
the anaesthetist to deliver good quality anaesthesia in a non-paralysed
patient. Satisfactory anaesthesia requires:
1. adequate cardiovascular and respiratory
stability
2. no or minimal patient movement
3. no awareness or recall of events during
the procedure
The monitor should function with all types of anaesthetics with the
possible exception of ketamine anaesthesia which is recognised to produce a
distinct form of anaesthesia from other agents.
The ultimate proof for a measure of anaesthetic depth is that it should
be capable of controlling automatically the delivery of an anaesthetic agent to
produce satisfactory anaesthesia in a patient breathing spontaneously during
surgery. Closed-loop anaesthesia (CLAN) systems have been developed using blood
pressure5or median frequency of the
compressed spectral array.6 However, these CLAN systems
have only been used in paralysed patients during surgery and required
intervention often because of unsatisfactory conditions. A system based on the
bispectral index has been reported to have successfully controlled the level of
sedation in 10 patients during spinal anaesthesia for elective orthopaedic
surgery. 7
A closed-loop control system based on the auditory evoked response has
been used to control the intravenous administration of propofol in 100 patients
breathing spontaneously and also in patients who received paralysing drugs
during surgery.8; 9 The quality of anaesthesia was judged to be satisfactory as
assessed by scores of autonomic activity, cardiovascular stability and minimal
movement during surgery. Patients were visited postoperatively and there was no
occurrence of awareness during the surgical procedures in any patient.
CLAN techniques have demonstrated the interdependence of hypnosis,
analgesia and stimulation and have illustrated that there is no single
concentration of an anaesthetic agent which results in satisfactory anaesthesia
for all patients. Indeed within individual patients, the requirements for
anaesthetic vary considerably depending on the degree of surgical stimulation
and the quality of analgesia provided at any point in time.
Relatively high concentrations of a hypnotic agent are required when a
low dose analgesic technique is used. When a high concentration of analgesic is
administered, smaller doses of hypnotic are then required to maintain
satisfactory. This is shown most clearly when patients have good premedication
and a fully functioning local block in place. In this situation, the
requirements for anaesthetic can be reduced to very low values.
Conclusion
The requirements for a depth of anaesthesia monitor are summarised in
the table. We can expect considerable progress in the developments of systems
to provide some estimates of the depth of anaesthesia. It is clear, however,
that anaesthetists must attempt to define the state of anaesthesia and lay down
conditions which these devices must meet.10
1. Russell
IF: Midazolam-alfentanil: an anaesthetic? An investigation using the isolated
forearm technique. Br.J.Anaesth.
1993; 70: 42-6
2. Davies
FW, Mantzaridis H, Fisher AC, Kenny GN: Middle latency auditory evoked
potentials during repeated transitions from consciousness to
unconsciousness. Anaesthesia
1996; 51: 107-13
3. Mantzaridis
H, Kenny GN: Auditory evoked potential index: a quantitative measure of changes
in auditory evoked potentials during general anaesthesia. Anaesthesia 1997; 52: 1030-6
4. Gajraj
RJ, Doi M, Mantzaridis H, Kenny GN: Analysis of the EEG bispectrum, auditory
evoked potentials and the EEG power spectrum during repeated transitions from
consciousness to unconsciousness. Br.J.Anaesth.
1998; 80: 46-52
5. Robb
HM, Asbury AJ, Gray WM, Linkens DA: Towards a standardized anaesthetic state
using isoflurane and morphine. Br.J.Anaesth.
1993; 71: 366-9
6. Schwilden
H, Stoeckel H: Closed-loop feedback controlled administration of alfentanil
during alfentanil-nitrous oxide anaesthesia.
Br.J.Anaesth. 1993; 70: 389-93
7. Mortier
E, Struys M, De ST, Versichelen L, Rolly G: Closed-loop controlled
administration of propofol using bispectral analysis. Anaesthesia 1998; 53: 749-54
8. Kenny
GN, McFadzean WA, Mantzaridis H: Propofol requirements during closed-loop
anesthesia. Anesthesiol. 1993; 79:
A329-A329
9. Kenny
GN, Mantzaridis H: Closed-loop control of propofol anaesthesia. Br.J.Anaesth. 1999; 83 (2):
223-8
10. Kenny
GN, Mantzaridis H, Fisher AC: Validation of anesthetic depth by closed-loop
control, Memory and Awareness in Anesthesia. Edited by Sebel PS, Bonke B,
Winograd E. Englewood Cliffs, Prentice Hall, 1993, pp 225-264.