Central Respiratory Effects
Of Benzodiazepines

John Loadsman - 13/2/96

These notes were written for a lecture given to department trainees. The information contained herein, including the figures, has been adapted from sources listed in the bibliography.






Drugs may effect the function of the CNS therefore in 2 ways generally:

  1. Interaction with interneuronal signalling
  2. Interaction with ion transport

NOTE: Drugs falling into Group 1 may have a very specific mechanism of action (e.g., Benzodiozepines) but their consequences are frequently diverse c.f. agents having specific group 2 effects which can have very specific consequences.

Benzodiazepine Receptor (BZR) ligands, especially the classic Benzodiazepines (BZs), are among the most specific centrally acting drugs falling into the first group with respect to

  1. Site - an allosteric modulatory site on the GABAA receptor
  2. Mechanism - interaction with signal recognition and transduction by that receptor



GABAergic neurones

GABA receptors

* BZs act only on GABAA receptors *



Common agonist



Selective agonist



Selective antagonist


Delta-Amino-valeric acid


Cl- channel

Ca++ channel (inhib)
ACh (inhib)
K+ channel (stim)

Allosteric modulation

Some convulsants

The GABAA receptor

* This forms the basis of the inhibitory function of the GABAergic system. *

Some examples of how GABAergic inhibition is thought to work:

The significance of GABAergic inhibition can be demonstrated by blocking it e.g. with GABA receptor blocking agents :


1975 - First suggestion that classic BZs (agonists) acted by enhancing GABAergic inhibition.

Since shown that BZ agonist facilitation of GABAergic transmission occurs in all areas of the CNS where GABAA receptors have been found.

BZR distribution has been shown to coincide with GABA receptor distribution by autoradiograms of animal brain sections after treatment with 3H-diazepam and more recently by P.E.T. scanning of live animal and human brains after injection with 11C-Flumazenil.

Radioisotope and Flourescent BZR bigands have also been used to demonstrate their high affinity and high specificity for the BZR and also the fact that these sites are saturable.

Studies of the action of BZR agonists on the GABAergic system have shown the following :

In 1985 the BZR was isolated using monoclonal antibodies and it was found that as well as BZs it bound GABA, leading to the conclusion that the GABAA receptor and the BZR were both part of one complex, the GABAA - BZR - Cl- channel complex.

This was comfirmed in 1987 when DNA for alpha and beta subunits of the complex were cloned and injected into cells not normally having GABAA receptors, resulting in the synthesis and membrane inclusion of fully functional GABAA - BZR - Cl- channel complexes. The gCl- of the cells varied in the predicted manner in response to GABA before and after addition of BZs .(No response was seen with addition of BZs alone). This confirmed the theory that the BZR is an allosteric modulatory site on the GABAA receptor and that GABA and BZs allosterically modulate each other's binding to their respective receptors.

It was thought in 1987 that the receptor complex contained 2 a subunits (containing the BZRs) and 2 b subunits (containing GABAA receptors) but by late 1990 subunits of 4 classes (alpha ,beta , gamma , delta) had been identified and coded with even further subdivision within the classes. Diverse structural and pharmacological heterogeneity of the complex has now been demonstrated.

It has also been shown in the 2nd half of the 1980's that the GABAA receptor complex occurs in virtually every neuron in the brain, on glia, and also exists in the periphery.



The core of all BZR ligands is a benzene ring joined to a 7-membered 1,4-diazepine ring:

The various substituents determine the ligand's pharmacological properties:

THE BZR AGONISTS (The Classic Benzodiazepines)




all produce the same effects ( to varying degrees )

  1. Anticonvulsant (at 20-25% occupancy)
  2. Anxiolytic (20-30%)
  3. Sedative / Hypnotic (25-30%/60-90%)
  4. Muscle relaxant (centrally mediated)
  5. Amnesic


MIDAZOLAM: an agonist - intrinsic efficacy +1.0

Ro 16-6028: a partial agonist - intrinsic efficacy +0.5

FLUMAZENIL: an antagonist - intrinsic efficacy 0

Ro 15-4513: a partial inverse agonist - intrinsic efficacy -0.5

Ro 19-4603: an inverse agonist - intrinsic efficacy -1.0




The first studies of the respiratory effects of diazepam were carried out in the late '60s. They examined CO2 responses only and most concluded there was no effect.

In 1971, Catchlove & Kafer used Read's rebreathing technique for the first time to examine the effect of 0.14mg/kg (10mg/70kg) i.v. diazepam on CO2 responses and steady-state gas exchange. VT decreased 20%, PaCO2 increased 19%, and VD/VT increased 34%. The slope of the CO2 response was depressed in 46%, unchanged in 23% and increased in the rest. 84% showed a rightward shift.

Forster et al (1980) studied deltaVE/deltaPETCO2 and deltaP0.1/deltaPETCO2 after 0.3mg/kg diazepam and 0.15mg/kg midazolam i.v.. Both drugs depressed both variables to a similar degree. No rightward shift could be demonstrated.

Gross et al (1982) showed the CO2 response to be depressed significantly for at least 25 minutes after 0.4mg/kg i.v. diazepam. The degree of depression correlated well with the level of sedation. Again no right shift demonstrated.

Power et al (1983) claimed neither 0.15mg/kg diazepam nor 0.075mg/kg midazolam i.v. caused depression of the CO2 response despite the mean slopes falling to 0.72 and 0.60 of their control values respectively. Main problem was n=7 (i.e. Power had no power! :-)

Alexander and Gross (1988) demonstrated a 53% decrease in the slope of the isohypercapnic (PETCO2 50mmHg) hypoxic response (deltaVE/-deltaSaO2) after 0.1mg/kg midazolam i.v. It also obliterated the haemodynamic response to hypoxia.

Gillis et al (1988) performed an elegant study where they applied various BZR ligands to the intermediate area of the ventral surface of the medula of cats anaesthetised with a-chloralose. Both midazolam (0.75-250mg/side) and chlordiazepoxide (100-1000mg/side) caused dose related depression of VE, VT, heart rate and blood pressure, although some tachyphylaxis was demonstrated. Respiratory frequency did not change. The effect was evident within 1 minute and lasted 30-60 minutes. Application of the agents to the rostral area caused neither effect and application to the caudal area caused hypotension without depressing respiration. Only the largest of these doses given i.v. had any effect but increasing i.v. doses (up to 1.0mg/kg midazolam) produced the same dose related response as the smaller topically applied ones, with the addition of depressed resiratory rate. Topically applied flumazenil (250mg/side) and bicuculline (a selective GABAA antagonist - 10mg/side) had no effect on their own but both prevented (if applied before) and reversed (if applied after) the depression caused by midazolam whether it was given topically or i.v. Ethyl-beta-carboline-3-carboxylate (a BZR inverse agonist - 12.5-30mg/side) applied topically to the medulla caused the opposite effects to midazolam and was also reversed by flumazenil.

Also in 1988, Skatrud et al examined the effects of up to 3 times the normal oral dose of triazolam. They claimed no effect on VE (VT decreased and f increased). However, they had a similar result with 0.15mg/kg i.v. morphine (a fairly big dose) and their intersubject variability was very large.

In 1990, Wettstein, Teeple and Morse examined the effects of a variety of BZR ligands and other agents (given i.v.) on the respiration of awake rhesus monkeys. The BZ agonists alprazolam (0.01-1.0mg/kg), lorazepam (0.3-10.0mg/kg) and quazepam (1.0-5.6mg/kg) all depressed VT and VE during both 5%CO2 and air inhalation. Pentobarbital (3.0-30mg/kg) also depressed frequency and it's dose/effect relationship on VT was steeper. Two beta-carbolines (beta-CCE (0.3-5.6mg/kg) and FG 7142 10mg/kg)) increased frequency and VE. Both flumazenil and CGS 8216 (a weak inverse agonist)(both 1.0mg/kg) reversed the effects of the agonists. Alprazolam reversed the stimulation of FG 7142.

Dahan and Ward claimed in 1991 that midazolam (0.025mg/kg followed by infusion at 1.0mg/70kg/h - a piddling dose!) did not reduce the acute ventilatory response to isohypercapnic hypoxia. In their discussion however they state there was a reduction but it failed to reach significance, hardly surprising when there were only five subjects and the PETCO2 was actually higher in the midazolam experiments! The midazolam also caused a greater hypoxic ventilary decline.

Gross, Weller and Conard reversed midazolam (0.13mg/kg) with flumazenil (1.0mg/kg) and found that the sedation and right shift of the CO2 response were rapidly reversed but the depression of the slope of the CO2 response was not.

Blouin et al (1993) confirmed midazolam's (0.12mg/kg) depression of the isohypercapnic hypoxic ventilatory response (deltaVE/deltaSpO2 reduced to 59%). This was fully reversed with flumazenil (1.0mg/kg).


BZR agonists almost certainly cause dose-related centrally mediated (medullary) respiratory depression. VT seems to be affected more than frequency and the sites of these actions may be different. Inverse agonists cause the opposite effect and pure antagonists reverse the effect. Flumazenil may not be a pure antagonist given the variability in it's reversal of the effects of agonists on ventilatory responses. Alternatively, the level of receptor occupancy may play a role in these differences.


  1. Wood & Wood: Drugs and Anaesthesia. 2nd Edition, 1990.
  2. Amrein et al: Clinical pharmacology of Dormicum (Midazolam) and Anexate (Flumazenil). Resuscitation, 16, Suppl (1988) S5-S27.
  3. W.Haefely: Pharmacology of the Benzodiazepine receptor. European archives of Psychiatry and Neurological Sciences, (1989) 238: 294-301.
  4. S.Ymer et al : Structural and functional characterisation of the g1, subunit of GABAA/benzodiazepine receptors. The EMBO Journal, Vol 9; No 10 (1990) : 3261-3267.