Background: Opioid-induced respiratory depression (OIRD) is a critical concern in the perioperative period and, more recently, for population health due to illicit consumption. Benzodiazepines are frequently co-administered in the clinical setting and are also frequently found in victims of opioid overdoses. The magnitude of the benzodiazepine effect and its potential to exacerbate opioid effects has not yet been evaluated at the level of preBötzinger complex (preBötC) neurons in vivo. Methods: The study was approved by the local Animal Care Committee and conformed to NIH standards. Adult New Zealand White rabbits (2.5-3.5kg) of either sex were anesthetized, tracheotomized, ventilated, decerebrated and vagotomized. Phrenic nerve activity was recorded from the C5 rootlet, time averaged and used to calculate inspiratory (TI) and expiratory (TE) duration and peak phrenic activity (PPA) (LabChart 7pro, ADInstruments, Colorado Springs, CO). The preBötC was functionally identified by stereotaxic coordinates and the tachypneic response to electrical stimulation with a concentric bipolar electrode (FHC, Bowdoin, ME). Extracellular recordings of individual preBötC neurons were performed with a 64-channel multielectrode array (Neuronexus, Ann Arbor, MI). Neuronal discharge activity (Fn) was quantified using cycle-triggered histograms (Spike2, CED, Cambridge, UK). Protocol: Neuronal recordings were obtained at baseline (PCO2 50 mmHg, FiO2 0.6), during IV infusion of the ultra-short benzodiazepine remimazolam at maximally effective dose-rate, and at near-apnea during additional IV infusion of the ultra-short mu-opioid agonist remifentanil. Both infusions were then discontinued, and neuronal recordings continued to confirm recovery of Fn. Multiple protocols were performed in one animal if possible. Neurons were pooled for analysis according to discharge pattern. Statistical comparison with Friedman RM ANOVA, data are median (25-75% range). All described effects are significant at p<0.05. Results: In 11 protocols in 6 rabbits, we recorded a total of 477 neurons. Remimazolam decreased respiratory rate through an increase in TI (0.1 (0.1-0.2)sec) and TE (0.5 (0.4,0.8)sec) while PPA decreased (-19 (-16 to -30)%). Remifentanil further increased TI (0.4 (0.3-0.7)sec) and TE (8.2 (6.7-14.4)sec) and decreased PPA (17 (14-31)%). An increase in dose-rates resulted in apnea. For every discharge type, in 29-53% of neurons, remimazolam decreased Fn (group A), in 23-53% increased Fn (group B), and in 12-40% did not change Fn (group C). A: Remimazolam depressed Fn consistently by ~50% in all neuron types; remifentanil caused an additional 10% decreased in Fn in pre-inspiratory (pre-I) neurons and there was an increase in I-parabolic (I-para) and non-respiratory modulated (NRM) neurons. B: Remimazolam increased Fn between 50-200%; remifentanil caused additional increase in I-decrementing and I-para neurons and no change in other neuron types. C: In inspiratory and NRM neurons where remimazolam did not change Fn, remifentanil increased Fn between 12-60%, but there was no effect on expiratory neurons. Conclusion: Previous work showed that remifentanil decreased Fn in the majority of neurons of each neuron type, with greater depression at higher doses. While remimazolam changed global respiratory parameters similarly to remifentanil, more than 50% of neurons had no change or an increase in Fn. Only a small subset of pre-I neurons showed an additional decrease in Fn. This points to pre-I neurons as the key to inspiratory on-switch and also highlights that “inhibition of inhibition” may play a significant role in the effect balance of respiratory depressant drugs. This abstract was presented at the American Physiology Summit 2026 and is only available in HTML format. There is no downloadable file or PDF version. The Physiology editorial board was not involved in the peer review process.
Feng et al. (2026) studied this question.