SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Intravenous lipid emulsion has proven benefit in lipophilic drug-induced cardiotoxicity. Its effect in reversal of central nervous system depression secondary to overdose with lipophilic psychotropic agents remains uncertain. Twenty adult New Zealand White rabbits were anaesthetised with 20 mg.kg−1 thiopental and randomised to receive either 4 ml.kg−1 saline 0.9% or 4 ml.kg−1 lipid emulsion 20% immediately afterwards. Depth of anaesthesia was monitored using bispectral index (BIS) at 1-min intervals. Duration of anaesthesia was measured as time to regain the righting reflex (ability of the animal to right spontaneously from dorsal recumbency to sternal recumbency). The BIS was greater in the control group (p = 0.011). The greatest BIS differential was observed immediately following treatment (mean (SD) BIS 75.0 (9.5) for saline vs 58.6 (10.4) for lipid, 95% CI 5.75–27.1; p < 0.001). No difference was observed in duration of anaesthesia (mean (SD) 15.5 (0.8) min for saline vs 15.6 (0.7) min for lipid, p = 0.86). Lipid emulsion administration may serve to increase central nervous system depression in the early phase of lipophilic toxin distribution.

Central nervous system (CNS) depression secondary to self-poisoning is a common reason for admission to the intensive care unit. Few specific antidotes to such poisonings exist [1], with historical attempts at analeptic use proving either of unqualified efficacy or demonstrably harmful [2]. Standard management of drug-induced CNS depression entails close observation progressing to intubation of the trachea and invasive supportive care when required. The concept of a generic agent that might reverse the CNS depressant effects of an ingested toxin, and potentially alleviate the need for mechanical ventilation, is therefore an attractive one.

A large body of literature exists supporting the use of intravenous lipid emulsions as an antidote to cardiotoxicity secondary to lipophilic drug intoxication [3]. The most frequently cited biological mechanism for this benefit is sequestration of toxin away from its effect site to an expanded intravascular lipid phase (sink), and possible augmentation of redistribution and elimination [4]. Following the same rationale, use of lipid emulsions has been extended to attempts at ameliorating CNS depression arising from overdose of highly lipophilic psychotropic drugs with anecdotally reported success [5, 6]. Demonstration of a clear benefit (and conversely lack of harm) for lipid use in these clinical circumstances has, however, been inconsistent [7]. Furthermore, little evidence exists from animal studies to support such use.

Previous investigators have shown a decreased overall duration of barbiturate anaesthesia in lipid treated rodents compared with saline controls [8]. More recently, lipid infusion has also been demonstrated to reverse thiopental-induced respiratory depression as an isolated measure of CNS effect in rats [9]. To date, however, no study has been published to our knowledge quantifying the effect of lipid infusion on the depth of toxin-induced CNS depression.

Quantitative electroencephalographic analysis, including derivation of the bi spectral index (BIS), has been validated as a measure of depth of anaesthesia in both humans and animals [10, 11]. These data suggest that BIS may also afford an objective measure of the CNS pharmacodynamics of certain toxicological ingestions, specifically those that demonstrate a general anaesthetic-like effect. Accordingly, the present study was conducted to investigate the effect of lipid emulsion on both the magnitude (as measured by quantitative EEG) and the duration (as determined by clinical parameters) of CNS depression, in a rabbit model of thiopental-induced anaesthesia.

Methods

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

This trial was performed over a 1-week period at the Ruakura Animal Research Facility, Hamilton, New Zealand. The study protocol was reviewed and approved by the Ruakura Animal Ethics Committee. Twenty adult female New Zealand White rabbits were studied. Before utilisation animals were kept in gender specific enclosures, with free access to food and water, controlled temperature and humidity and standard 12-h day/night cycle, in accordance with institutional guidelines.

On the day of use, rabbits were weighed and randomly assigned to either control (saline 0.9%), or treatment (Intralipid® 20%; Baxter Healthcare Ltd, Auckland, New Zealand) groups. After placement in an individual restraining box (Iffa-Credo, L’Arbresle, France), all animals underwent venous cannulation of a marginal ear vein with saline-filled 20-G intravascular catheter. The central artery of the contralateral ear was similarly cannulated, and connected (Edwards Lifesciences Pressure Transducer, Irvine, CA, USA) to a Hewlett Packard 78834A neonatal monitor for continuous assessment of invasive arterial blood pressure. The pressure transducer was room-air zero-calibrated with reference to the mid thorax.

After subcutaneous administration of a total of 1 ml lignocaine 1%, four scalp-needle electrodes were placed in a configuration previously described for EEG monitoring in rabbits [11]. One electrode was placed 2 cm caudal to the lateral canthus of each eye, one reference electrode was placed in the midline overlying the frontal bone, and the ground electrode was placed 2 cm to the left of the central electrode. These were connected to an EEG monitor (A-1050; Aspect Medical Systems Inc, Natick, MA, USA). Low and high frequency filters were set to 2 and 70 Hz, respectively. The monitor was configured to display the unprocessed EEG tracing as well as the derived BIS value to allow correlation of the two by an observer, and potential detection of artefact.

After a 2-min period of stabilisation, baseline haemodynamic and respiratory metrics and EEG/BIS values were recorded. Animals were then given a total of 20 mg.kg−1 sodium thiopental (LINK Pharmaceuticals Ltd, Horsham, UK) in divided aliquots. Initial dosing of 10 mg.kg−1 (30 mg.ml−1 thiopental) was infused over 20 s (commencement of infusion nominally t = 0). Two subsequent doses of 5 mg.kg−1 thiopental of identical solution were administered over a similar period at t = 1 min, and t = 2 min, respectively. This anaesthetic regimen was adopted following initial dose-ranging experiments (= 3) that demonstrated maintenance of spontaneous ventilation and cardiovascular stability, unlike animals given 20 mg.kg−1 as a single bolus, which demonstrated profound respiratory depression.

At 30 s following the third dose of thiopental (t = 2.30 min), animals in the control group received 4 ml.kg−1 saline 0.9% over 30 s. Animals in the intravenous lipid emulsion group received 4 ml.kg−1 Intralipid 20% over a similar period.

Heart rate, respiratory rate, mean arterial pressure and the raw BIS values were recorded at baseline, and subsequently at 1-min intervals directly to a standardised data collection template. The presence or absence of reflexes, indicating depth of anaesthesia, were additionally recorded at 1-min intervals. These included the lash reflex, corneal reflex (corneal wipe with cotton wool) and response to pain (as defined by any movement in response to a standardised noxious stimulus (ear pinch with a household clothes-peg distant from vascular access sites)). Time to loss of any reflex, and consistent return of all three reflexes (presence of all three reflexes for three consecutive minutes), were recorded. Assessment of all reflexes was performed by the same investigator throughout the study, who was blinded to both the study drug and the EEG signal. A second un blinded investigator recorded values from the monitoring equipment, observed the unprocessed EEG and administered study agents.

At 15 min, final observations were made and the animals were removed from the restraining box to a separate enclosure. Time from induction of anaesthesia to return of the righting reflex, and time to standing (ability to lift the trunk against gravity), were then recorded. Following recovery, all animals were returned to the original enclosure and utilised at a later date in a second study.

Sample size estimation was based on previous work outlining the duration of thiopental-induced anaesthesia in rabbits [12]. This suggested a 1.4 SD difference in arousal time to be adequately detected with n = 10 for each experimental group, with power set at 80% at the 0.05 level of significance. Statistical analysis of all variables was conducted using graphpad Prism (version 5.0 GraphPad Software Inc, La Jolla, CA, USA). Comparison of continuous variables was conducted using the two-tailed Student’s t-test following assessment for normality with the Kolmogorov–Smirnov statistic. Continuous variables were compared across time with saline group as control, by two-way repeated measures ANOVA with Bonferroni post-testing when significance was achieved (p < 0.05). Delta BIS (defined as change in BIS score across time, computed at 1-min intervals) was additionally compared as a post-hoc variable of interest. A two-tailed p < 0.05 was considered statistically significant.

Results

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Baseline animal characteristics are presented in Table 1. BIS was greater in the saline-treated group (p = 0.011) Figure 1. The greatest BIS differential was observed at t = 4 min (mean (SD) BIS 75.0 (9.5) for saline vs 58.6 (10.4) for lipid, 95% CI for difference 5.8–27.1; p < 0.001). A significant difference in delta BIS was observed at t = 4 min only (0.8 (22.4) min−1 for saline vs −13.0 (10.7) min−1 for lipid, 95% CI for difference 3.1–24.5; p < 0.01).

Table 1.   Baseline animal characteristics. Values are mean (SD).
 Saline (n = 10)Lipid (n = 10)
  1. MAP, mean arterial pressure.

Age; days77 (5)74 (5)
Weight; g2385 (383)2711 (340)
Heart rate; beats.min−1214 (29)225 (19)
MAP; mmHg94 (10)90 (11)
Respiratory rate; breaths.min−190 (19)96 (15)
Bispectral index98 (0.7)98 (0.5)
image

Figure 1.  Bispectral index against time according to group. Dotted lines indicate thiopental administration. The solid bar indicates rescue therapy (saline or lipid emulsion) infusion. Values are mean (SD). inline image, lipid and inline image, saline.

Download figure to PowerPoint

Mean arterial pressure and respiratory rate are presented in Figs 2 and 3 respectively. No difference was observed in any of these variables across time between groups. No difference in the time to loss or recovery of any clinical reflex was observed between groups (Figs 4 and 5). No difference was observed in time to righting (mean (SD) 15.5 (0.8) min for saline vs 15.6 (0.7) min for lipid; p = 0.860), or time to weight bearing (15.8 (1.0) min for saline 15.7 (1.0) min for lipid; p = 0.827) between groups.

image

Figure 2.  Mean arterial pressure against time according to group. Values are mean (SD). inline image, lipid and inline image, saline.

Download figure to PowerPoint

image

Figure 3.  Respiratory rate against time according to group. Values are mean (SD). inline image, lipid and inline image, saline.

Download figure to PowerPoint

image

Figure 4.  Number of animals displaying intact reflexes vs time; lipid group. inline image, lash; inline image, cornea and inline image, pain.

Download figure to PowerPoint

image

Figure 5.  Number of animals displaying intact reflexes vs time; saline group. inline image, lash; inline image, cornea and inline image, pain.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

Intravenous lipid emulsion significantly increased the depth, but not the duration, of thiopental anaesthesia in this rabbit model. This suggests that lipid emulsion serves to increase peak brain thiopental concentration when administered immediately after induction. These findings superficially appear at odds with the postulated lipid sink as the main mechanism of action put forward for intravenous lipid emulsions. Simple application of the lipid sink theory might suggest that an expanded plasma lipid phase should mediate increased toxin washout from the target organ, the brain, as has been previously shown for bupivacaine in the rat isolated heart [13]. A consideration of the nature of thiopental CNS kinetics nevertheless yields potential explanations for this discord.

Following a bolus of thiopental, the offset of action results from drug redistribution to a rapid peripheral compartment (lean aqueous tissue, chiefly muscle) with later distribution to a slow (primarily adipose tissue) compartment [14–17]. Accordingly, multi-compartmental mathematical modelling may be used to explain the tri-exponential decline in plasma thiopental levels observed in humans and other species [15–17]. Consideration of such a construct provides potential explanations for our observed findings.

We propose that the creation of a high affinity lipid compartment within plasma serves to sequester lipophilic thiopental [18], thereby impeding thiopental off loading to the highly-perfused tissues of the rapid peripheral compartment. Sustained initial plasma thiopental levels would thus be available to equilibrate with the brain, resulting in deeper anaesthesia. Such decreased rapid inter-compartmental clearance has previously been demonstrated to result in lesser induction dose requirements [19, 20]. Lipid infusion might also be predicted to increase the equilibration time with the brain given the same rationale. While hypothetical, the action of lipid as a subway rather than a sink affords a rational explanation for our findings.

It is possible that lipid emulsions may act via some other mechanism, such as altering the constant keo affecting-equilibration between blood and brain. Changes in this variable for other highly lipophilic drugs have been shown to affect peak effect site concentrations, but not the subsequent rate of redistribution [21], with findings similar to ours. This change would, however, necessitate that lipid either altered brain perfusion or somehow increased rather than decreased the brain: blood partition coefficient. An independent direct effect of lipid emulsion on the brain is furthermore not excluded. Prior evidence would nevertheless suggest that the direction of any such effect would most likely be in the opposite direction to that observed. Previous investigators have found an increase in oxygen consumption of rat cortical slices incubated with lipid emulsion [8], and activation of N-methyl D-aspartate (NMDA) receptors in vitro following lipid exposure [22].

Measurement of duration of anaesthesia in the present work was limited due to the nature of animal restraint during the initial 15 min of the study protocol. However, inspection of the BIS-time curve and other reflexes in both groups suggests genuine differences in duration of anaesthesia to be unlikely. This observation differs from previous work. Russell and Westfall [8] found administration of lipid emulsion following thiopental induction to decrease the duration of anaesthesia in rats. Differences in thiopental and lipid emulsion dosing, in addition to inter-species variability, may in part explain the observed disparity between this work and those of our experiment. Rapid bolus dosing has been shown to increase the duration of thiopental anaesthesia in rabbits when compared with slower administration [23], such as the staggered dosing in our study. In addition, in a pilot study by Cave et al. [9] examining thiopental-induced respiratory depression in rats, lipid-treated animals showed greater improvements in respiratory rate compared with saline control. However, no measure of depth of anaesthesia was reported, and as in the current study, there were no significant differences reported in absolute respiratory rate at any time point between groups.

In the present study we have failed to obtain sequential estimates of plasma, or effect site, thiopental concentration. As such, any postulated alterations in drug pharmacokinetics following lipid application are hypothetical only. Nevertheless, EEG/BIS has been demonstrated to decrease with increasing plasma concentration of anaesthetic agents [24]. Similarly, another derived EEG value, spectral edge, has been shown to correlate with measured thiopental concentrations after induction [25]. In our model, BIS vs time curves were consistent with expected effect site concentrations following induction and redistribution. In addition, effect site concentrations are difficult to measure directly in vivo, and gross extracellular fluid estimation of anaesthetic agent concentration may not reflect receptor site drug concentration [26].

Our study is subject to a number of additional limitations. One investigator was un blinded to the study intervention, raising the possibility of observer bias. However, all collated measurements were recorded by this observer in an objective and systematic fashion. Rabbits remained spontaneously breathing throughout the protocol. In the absence of arterial blood gas measures, however, we are unable to comment on the potential for differences in ventilation-dependent variables (pH, Pco2, Po2) to have impacted on cerebral function and thereby derived BIS. There was, however, no significant difference observed in respiratory rate between the two groups.

Artefact in the EEG due to electromyographic interference in the range of frequencies monitored in this study may have biased reported BIS values [27]. Nystagmus, which might generate such an electromyographic potential, was frequently observed during emergence but its presence was not quantified for each group. Raw EEG waveforms were nevertheless continuously observed for presence of artefact. The model of CNS toxin administration utilised in this study presents additional caveats. Our staggered thiopental-dosing protocol may have impacted on the depth of anaesthesia, which was less than expected in control animals, potentially also altering the peak pharmacokinetic profiles.

Any clinical implications of these findings must be tempered by the methodological limitations of our work, in particular the rapid administration of toxin and brief delay to lipid infusion. In this scenario, lipid administration appears to have hindered initial drug redistribution to peripheral compartments. However, any clinical use of lipid as an antidote is invariably only possible following the patients’ presentation at some juncture distant from toxin ingestion, and at a time point when fast-equilibrating peripheral compartments may already be near-saturated with toxin. Distribution clearance of toxin at this point may depend more on the slow peripheral compartments [28] where the potential pharmacokinetic effects of lipid infusion are as yet un-studied.

Clinicians considering the use of lipid emulsion in the setting of reduced level of consciousness secondary to intoxication should recognise the possibility that there may be ongoing decline in level of consciousness. Preparation for rapidly securing the airway should be made concomitant with use of lipid emulsion.

Acknowledgements

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References

The authors wish to thank Mr Ric Broadhurst for his assistance in performance of study procedures. Funding for the present study was via a grant from the New Zealand Lotteries Commission. No competing interests declared.

References

  1. Top of page
  2. Summary
  3. Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References