Effects of fentanyl overdose-induced muscle rigidity and dexmedetomidine on respiratory mechanics and pulmonary gas exchange in sedated rats

Abstract

The objective of our study was to establish in sedated rats the consequences of high-dose fentanyl-induced acute muscle rigidity on the mechanical properties of the respiratory system and on the metabolic rate. Doses of fentanyl that we have previously shown to produce persistent rigidity of the muscles of the limbs and trunk in the rat (150–300 μg/kg iv), were administered in 23 volume-controlled mechanically ventilated and sedated rats. The effects of a low dose of the FDA-approved central α-2 agonist, dexmedetomidine (3 μg/kg iv), which has been suggested to oppose fentanyl-induced muscle rigidity, were determined after fentanyl administration. Fentanyl produced a significant decrease in compliance of the respiratory system (Crs) in all the rats that were studied. In 13 rats, an abrupt response occurred within 90 s, consisting of rapid rhythmic contractions of most skeletal muscles that were replaced by persistent tonic/tetanic contractions leading to a significant decrease of Crs (from 0.51 ± 0.11 mL/cmH₂O to 0.36 ± 0.08 mL/cmH₂O, 3 min after fentanyl injection). In the other 10 animals, Crs progressively decreased to 0.26 ± 0.06 mL/cmH₂O at 30 min. There was a significant rise in oxygen consumption (V̇o₂) during these muscle contractions (from 8.48 ± 4.31 to 11.29 ± 2.57 mL/min), which led to a significant hypoxemia, despite ventilation being held constant. Dexmedetomidine provoked a significant and rapid increase in Crs toward baseline levels, whereas decreasing the metabolic rate and restoring normoxemia. We propose that the changes in respiratory mechanics and metabolism produced by opioid-induced muscle rigidity contribute to fentanyl lethality.

NEW & NOTEWORTHY The decrease in respiratory compliance and increased metabolism-induced hypoxemia produced by an overdose of fentanyl, in and of themselves, contribute to fentanyl toxicity.

INTRODUCTION

Opioid-induced acute hypoventilation is responsible for the life-threatening toxicity of opioids, as it can lead to a terminal hypoxic/anoxic cardiac arrest, by pulseless electrical activity or ventricular arrhythmia (1). Such a lethal outcome can result from a rapid intravenous injection of a high-dose opioid agonist, but can also occur at lower doses in patients suffering from chronic respiratory failure or obesity hypoventilation syndrome (2). As first described by Hamilton and Cullen in 1953 (3), intravenous opioid administration also generates abrupt and prolonged tetanic contractions of a large mass of skeletal muscles (4). Opioid-induced muscle rigidity is now well recognized in the operating room and in intensive care units alike, as a significant side effect of opioids (5) and is well documented with lipophilic synthetic opioids such as fentanyl or alfentanil, for instance (4). Opioid-induced muscle contractures/contractions can affect breathing as a result of the tonic contractions of respiratory inspiratory and expiratory muscles (6), including the abdominal muscles (4–7). These persistent tonic contractions of the muscles involved in respiration have a major impact on the dynamic and the mechanics of breathing, as they produce a reduction in chest wall compliance (and thus in total respiratory compliance), literally counteracting breathing movements (5, 8, 9). The terms “chest wall rigidity” and “wooden chest syndrome” have been coined to describe this effect, which has been characterized in various animal models (7) as well as in patients receiving intravenous fentanyl. Marked truncal and abdominal muscle rigidity along with dyskinesia have well been documented during procedural sedation or at supervised injection sites (4, 5, 10). For instance, Kinshella et al. (10) have reported, at a supervised injection site, a range of muscle contractures from jaw clenching to decorticate posturing with arms bent in toward the body, legs held out straight, clenched fists, and overall stiffness. Muscle rigidity was the most common type (almost half) of atypical overdose presentation. Isolated torso rigidity or rigid posturing was also frequently observed (10). These impressive symptoms seem to progress extremely rapidly as central nervous system and respiratory depression develop.

Of interest, opioid-induced muscle rigidity and dyskinesias appear to be produced by the inhibition of neurons, which normally exert a tonic suppression of adrenergic neurons in the locus coeruleus (11–16). This effect, therefore, involves neuronal networks different from those causing a direct depression in respiratory drive. The role of the activation of adrenergic neurons as a mechanism of chest wall rigidity is supported by the rapid beneficial effect resulting from the administration of central acting α-2 agonist agents that have been shown to reduce skeletal muscle rigidity (15, 17).

Another potentially deleterious effect of opioid-induced chest wall rigidity, which has been overlooked, is its direct impact on the alveolar gas equations for O₂ and CO₂. Since the alveolar partial pressures in O₂ and CO₂ are dictated by the ratio between O₂ uptake or CO₂ production rate and alveolar ventilation, any increase in metabolism resulting from these impressive muscle contractions could magnify the decrease in alveolar Po₂ and increase in alveolar Pco₂ otherwise produced by opioids-induced hypoventilation (see physiological principles in the methods).

The objective of the present study was to determine, in adult rats, the time course and the amplitude of the changes in respiratory mechanics produced by a unique acute administration of fentanyl, at doses that we have previously established to produce a potentially lethal apnea, along with muscle rigidity (18, 19). We have also investigated the effects of fentanyl-induced muscle rigidity on the level of metabolic rate and arterial oxygenation. Finally, we have tried to determine the effects of the central acting α-2 agonist, dexmedetomidine, on the changes in respiratory mechanics and metabolism produced by fentanyl. This agent, which is already FDA approved (20), has been shown to oppose opioid-induced muscle rigidity (15, 17, 21), but its potential benefit on blood gas homeostasis remains to be shown.

Our hypothesis is that the changes in respiratory compliance and V̇o₂ contribute to the respiratory failure produced by an opioid overdose and the centrally acting α-2 agonist has a rescuing effect on both respiratory mechanics and blood gas homeostasis.

METHODS

Animal Model

A total of 48 adult male rats (Sprague–Dawley, Charles River) weighing 531 ± 114 g (13–15 wk of age) were studied. Fifteen of these rats (490 ± 130 g) were studied to determine the optimal doses of dexmedetomidine (see paragraph below) to be used in our experiments, whereas the 23 remaining rats were part of the actual study. The Penn State College of Medicine Institutional Animal Care and Use Committee approved the study.

The rats were housed at the Animal Resource Services at the Penn State College of Medicine, which conforms to the requirements of the US Department of Agriculture and the Department of Health, Education and Welfare. Rats were free of pain and were provided with food and water ad libitum, on a standard 12 h (7:00 AM to 7:00 PM) light/dark schedule, under the direct supervision of veterinarians. The Penn State College of Medicine is accredited by the American Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC), Animal Welfare Assurance Number A3045-01. Members of the full-time veterinary staff of the Department of Comparative Medicine are available 24 h/day, 7 days/wk.

The 23 rats used in the study were initially anesthetized with inhalation of 3%–5% isoflurane in O₂, followed by an intraperitoneal injection of ketamine (100 mg/kg ip). Isoflurane inhalation was then stopped and anesthesia was maintained by successive intravenous (iv) injections of ketamine (5–10 mg/kg iv bolus, using a tail vein catheter 24 G-DB), as needed, to maintain the animals unresponsive to any external stimuli and until the administration of fentanyl overdose. The eyes were lubricated using sterile eye lubricant (Refresh) during the experiment. The anesthetic depth was evaluated by monitoring the response to toe pinch during the protocol every 10 min. If the animal responded to the toe pinch (motor response and changes in breathing pattern), the procedure was stopped temporarily, and a new injection of ketamine was performed. This approach was used if similar reactions were present during the catheter placement or the tracheostomy procedure until the large dose of fentanyl was administered.

Rats were tracheostomized (14 G surflo catheter), and the tracheostomy was attached to a pneumotachograph (Series 1100, Hans Rudolph, Inc. Shawnee, KS) to determine the respiratory flow. A catheter (PE-50) was placed into the right carotid artery to continuously monitor arterial blood pressure (ABP, via a pressure transducer MLT844, AD Instruments, Colorado Springs, CO) and for sampling arterial blood. Two similar catheters were placed into one jugular and one femoral veins or two jugular veins for fentanyl and dexmedetomidine injections. Animals were mechanically ventilated in volume control mode (CWE, SAR-830/AP, Ardmore, PA; Fig. 1). Inspiratory and expiratory flows were measured (22, 23), and tracheal pressure (trans-thoraco-pulmonary pressure, Prs) was constantly monitored as shown in Fig. 1. Mixed expired O₂ and CO₂ fractions were continuously determined, as previously described (18, 19), using fast responding O₂ (Oxystar-100, CWE Inc. Ardmore, PA) and CO₂ (model 17630, VacuMed, Ventura, CA) analyzers. Body temperature was monitored with a rectal probe (Thermalert TH-5, Physitemp, Clifton, NJ) and maintained at around 37°C using a lamp and heating pad. At the end of the experiment, rats were euthanized by a lethal injection of high-dose barbiturate intravenously (Euthasol).

Figure 1.

A: illustration of the setting used for the measurement of flow and Prs (trans-thoraco-pulmonary pressure, determined at the trachea) in mechanically ventilated rats for the computation of the compliance of the respiratory system (Crs), lower respiratory resistance (Rrs), the passive time constant of the respiratory system (τrs) along with V̇o₂ (see methods for more details). All signals were digitized at 400 Hz, tidal volume was determined by integration of flow signal, as depicted in B, while V̇o₂ was computed every 5 s. Note, in B, flow is constant during the inspiratory phase (volume-control ventilation); as a result, volume [the integral of V̇(t)] increases linearly with time along with Prs(t), with a relationship that is dictated by Crs. Neglecting the inertial properties of the respiratory system, the offset in Prs (immediate increase in Prs when flow rises abruptly, with trivial changes in volume) corresponds to the resistive component of the total pressure. The resistance of the equipment (tracheal tube and connectors distally to pressure measurement) was determined by establishing the pressure-flow relationship at different flows (C), the resistive properties of tracheal tube and connectors were better described by a relationship: P = K₁.V̇+ K₂.V̇². V̇o_2, oxygen consumption.

All signals were digitized at 400 Hz using an analog-to-digital data acquisition system (Power Lab 16/35, AD Instruments, Inc. Colorado Springs, CO), visualized online, and stored for further analysis by LabChart8 (AD Instruments, Inc. Colorado Springs, CO). Mixed expired O₂ and CO₂ fractions, and flow and tracheal pressure signals were used for the computation of minute ventilation, O₂ consumption (V̇o₂) along with the monitoring of arterial blood pressure and heart rate. Blood gases were measured using a Vetscan i-STAT1 analyzer (ABAXIS, Union City, CA).

The resistances of intratracheal catheter and connectors were determined as the drop in pressure produced by the equipment in response to different levels of constant flows of air (high precision rotameter), corresponding to the physiological range of instantaneous flows (0–20 mL/s). Although the flow-volume relationship of the intratracheal catheter and connectors was better described as P = K₁. V̇(t) + K₂$\dot{V}$(t)² (Fig. 1), the inspiratory resistances of the equipment were estimated as a slope of the linear portion of pressure-flow relationship for the sake of simplicity.

Principles Used in the Present Study to Determine the Passive Mechanical Properties of the Respiratory System

As illustrated in Fig. 1, animals were ventilated using a volume-controlled mode. Unfortunately, the ventilator we are using to ventilate our rats does not allow, at a frequency of 70 breaths/min, to generate a pause at the end of inspiration. The determination of a drop in Prs corresponding to the resistive component of the pressure was therefore not possible. Some of the principles described in the following paragraphs can be found in the following references (24–28).

As flow is constant during inspiration (28), any change in respiratory compliance will translate into a change in driving pressure generated by the ventilator to maintain a constant flow (or a linear increase in volume over time). Based on the flow and pressure signals, the rats did not display any inspiratory movements during all of the measurements, the compliance of the respiratory system (Crs) was estimated as the slope of the pressure-volume relationship during the period of inspiration when no change in flow occurs [V̇(t) = V̇ss], discarding the first 50 ms of inspiration (Fig. 2). Indeed, in these conditions, the relationship between the trans-thoraco-pulmonary pressure (Prs) and volume (V) as a function of time (t) can be written as:

$$\text{Prs}\left(t\right)=\text{V}\left(t\right)\text{/Crs+ Po}$$ (1)

where Po is the sum of the resistive component of the respiratory system [Rrs. V̇(t)] and any positive end-expiratory pressure (PEEP), neglecting the inertial properties of the respiratory system:

$$\text{Po}=\text{Rrs}.\dot{\text{V}}\text{ss}+\text{PEEP}$$ (2)

Figure 2.

A: typical example of a recording of respiratory flow ($\dot{V}$), tidal volume (VT), and trans-thoraco-pulmonary pressure (Prs) in a mechanically ventilated rat, displaying the portion of the volume and pressure signals that were used to compute Crs, discarding the first 50 ms as shown in B. B: individual data used to compute the linear regression between pressure and volume; as Prs (t) = V (t)/Crs+ Po, Crs was determined as the slope of the pressure-volume relationship during inspiration. C: different analysis, applied to the same data, where Crs was obtained by multilinear regression analysis over the entire period of inspiration. In contrast to B, this analysis is performed using the flow, the volume, and the pressure signals starting at the very onset of inspiration: Prs (t) = V (t)/Crs+ Rrs. $\dot{V}$(t). Using this method, Crs and Rrs could be directly estimated. Prs(t) signal is shown on the left-hand side of the panel, whereas Prs(t) reconstructed from Crs and Rrs computed values are shown on the right-hand side. D: comparison of Crs values obtained by the two methods using a total of 78 values that were obtained at the same time before and after fentanyl in 13 animals (3 pairs of data). The approach shown in B was used for continuous (breath-by-breath) estimation of Crs, whereas the values Crs and Rrs computed by multiple linear regression were used to compare specific time points. Crs, compliance of the respiratory system; Rrs, lower respiratory resistance.

Since the flow is constant during the inspiration [V̇(t) = V̇ss] and no extrinsic PEEP was applied, Rrs can be estimated as a first approximation as Po/V̇ss (after subtraction of the resistance of the intratracheal catheter and connectors). Of note, the resistive component of Prs, i.e., Rrs.V̇ss, also corresponds to the offset of the pressure signal at the very onset of inspiration, when the flow reaches its steady state with trivial changes in volume. This analysis was automatically applied to every breath and averaged every 15 s. This automatic computation of Crs and Rrs was compared with their “manual” determination, at a specific timing, obtained by multilinear regression analysis (analysis using Microsoft Excel data analysis) over the entire period of inspiration. A multiple regression analysis (29, 30) was performed using the flow, volume, and pressure signals from the very onset of inspiration, as follows:

$$\text{Prs}\left(t\right)=\text{V}\left(t\right)\text{/Crs+}\dot{\mathrm{V}}\left(t\right)·\text{Rrs\hspace{0.17em}}(+\text{PEEP})$$ (3)

Using this method, Crs and Rrs can therefore be directly estimated. The correlation between the two approaches (Eqs. 1 and 3) was established using 39 pairs of data (78 values) determined at the same time (6 values/animal in 13 rats). The two methods led to similar values of Crs (Fig. 2). Equations 1 and 2 were used throughout the study to monitor Crs “online,” whereas Eq. 3, which was not possible to compute “automatically” on a breath-by-breath basis, was used for comparisons at specific time points. Finally, the passive time constant of the respiratory system (τrs) was determined as the slope of the linear part of the expiratory flow-volume curve (Fig. 3):

$$\text{V}\left(t\right)=\text{τrs}·\dot{\mathrm{V}}\left(t\right)+{\text{V}}_0$$ (4)

where V₀ correspond to the lung volume actively maintained above the functional residual capacity (FRC, extrapolation of the $\dot{V}$-V relationship to zero flow). Since the resistances of the expiratory circuit are practically impossible to determine (as they include part of the circuitry of the ventilator), only the changes in τrs produced by fentanyl and treatment (saline vs. dexmedetomidine) were considered.

Figure 3.

Example of a recording of the change in flow and volume during expiration. The passive time constant of the respiratory system (τrs) was determined as the slope of the linear part of the expiratory flow-volume curve: V(t) = τrs.$\dot{V}$(t) +V₀ (see text for more details). Of note, since the resistances of the expiratory circuit are practically impossible to determine, only the differences in τrs produced by fentanyl and the treatment (saline vs. dexmedetomidine) were considered in text.

Effects of Rigidity-Induced Increase in Metabolism on Alveolar Po₂ and Pco₂

The alveolar gas equations for O₂ predicts that the level of alveolar partial pressure in O₂ ($\text{P}{\text{A}}_{{\text{O}}_2}$) is dictated by the ratio between oxygen consumption (uptake; V̇o₂) and alveolar ventilation (V̇A; 31, 32):

$$\text{P}{\text{A}}_{{\text{O}}_2}=\text{P}{\text{I}}_{{\text{O}}_2}-[K(\dot{\text{V}}{\text{o}}_2/\dot{\text{V}}\text{A})]$$ (5)

where $\text{P}{\text{I}}_{{\text{O}}_2}$ is the partial pressure of O₂ in the inspired gas.

A similar relationship can be written between the alveolar partial pressure in CO₂ ($\text{P}{\text{A}}_{\text{C}{\text{O}}_2}$), CO₂ output (V̇co₂), and V̇A:

$$\text{P}{\text{A}}_{\text{C}{\text{O}}_2}=K\dot{\text{V}}{\text{co}}_2\text{/}\dot{\text{V}}\text{A}$$ (6)

Our rationale is that fentanyl-induced muscle rigidity should increase V̇o₂ and V̇co₂; this rise in pulmonary gas exchange rate should according to Eqs. 5 and 6 decrease $\text{P}{\text{A}}_{{\text{O}}_2}$ and $\text{P}{\text{A}}_{\text{C}{\text{O}}_2}$ at any given level of ventilation, regardless of, or in addition to, any other mechanisms that could lead to a reduction in V̇A. We tried to characterize the contribution of an increase in metabolism and thus in pulmonary gas exchange produced by fentanyl-induced muscle rigidity on the development of hypoxemia.

Rationale for the Choice of the Doses of Fentanyl and Dexmedetomidine

We have previously determined the relationships in unsedated rats between breathing pattern (including apnea duration), minute ventilation, metabolism, and fentanyl dose (18, 19). The doses from 50, 100, and 300 μg/kg infused over a minute were previously shown to produce an immediate apnea, followed by a very long period of very severe hypoventilation; at the dose of 100 and 300 μg/kg very rapid and persistent tonic muscle contractions, including the limb, back, and abdominal muscles were present in ∼40% and 100% of the animals, respectively (19), and lasted for at least 1 h. We, therefore, chose the dose of fentanyl used in these studies (150 or 300 μg/kg over 1 min) to mimic the symptoms of an opioid overdose found in unsedated animals (apnea and rigidity; 18, 19).

For dexmedetomidine, we first determined, in three unsedated rats, the doses of dexmedetomidine that did not produce sedative effects of this agent. We found that 3, 4.5, and 6 μg/kg iv had no sedative effects, based on the lack of decrease in locomotor activity, metabolism, and responsiveness to an auditory stimulus. These three doses of dexmedetomidine would correspond to doses ranging from 3 to 6 μg/kg iv in an adult human, in keeping with the allometric relationship between humans and rats (33–35) and could be considered as low doses in humans (36, 37). In 12 rats (3 rats for each dose and 3 rats for saline), dexmedetomidine or saline was administered intravenously 3 min before 300 μg/kg fentanyl to determine the capacity of dexmedetomidine to reduce the initial rhythmic phasic muscular contractions produced during the period of apnea following the administration of these high doses of fentanyl in unsedated animals (18, 19). This motor activity was determined by visual inspection of the flow signal in the plethysmograph and the behavior of the rats while measuring breathing patterns, as previously described (18, 19). The motor response was characterized by two phases, akin to the effects we have previously reported (18, 19): an initial volleys of rhythmic skeletal muscle contractions produced almost immediately by fentanyl, which typically subsided, within 1–2 min (see also results for more details). During this phase, there were hyperextension and abduction of the four limbs along with a rigidity of the muscle of the trunk and back developed. This response was replaced by a tonic and continuous flexion of the upper and lower limbs and a contracture of the back muscles (from the cervical to the sacral regions including the tail) and of the abdominal and trunk muscles, as displayed in Fig. 4. We found that the duration of the initial phase, rhythmic bursts of contractions, was reduced, according to the dose of dexmedetomidine (3–6 μg/kg) when administered before fentanyl (Fig. 5). At the dose of 6 μg/kg iv, the startling response produced by fentanyl was virtually abolished and no persistent tonic rigidity developed. However, since this dose was associated to a prolongation of the duration of apnea produced by fentanyl, the dose of 3 μg/kg/min infused iv over 3 min was therefore selected for the present study.

Figure 4.

Example of the posture and muscular tone displayed by 2 unsedated control rats (here 1 and 9 min after fentanyl 300 μg/kg). First, an initial period of 1–2 min wherein the animals presented volleys of phasic and tonic contractions were observed (A). During this phase hyperextension and abduction of the limbs along with a rigidity of the muscle of the trunk and back developed. This response was replaced by a tonic and continuous flexion and adduction of the upper limbs and flexion the lower limbs, with persistent contraction of the back muscles (form the cervical to the sacral regions with an extension of the tail) along with clear tetanic rigidity of the abdominal muscles (B and C).

Figure 5.

A: effects of saline and dexmedetomidine (at 3 different doses, 3, 4.5, and 6 μg/kg over 3 min) in 12 unsedated rats (3 rats/dose and one group of 3 rats for saline). Dexmedetomidine or saline was administered 3 min before fentanyl overdose (300 μg/kg). The duration of the initial volleys of phasic muscle contractions—which typically lasted less than 1 min—along with the duration of the initial apnea, were compared. There was a clear reduction of the duration of the initial volley of phasic muscle contractions produced by fentanyl in keeping with the level of dexmedetomidine. At 6 μg/kg, however, apnea duration was significantly longer that in any other condition. B1: maximal changes in Prs occurring within 3 min following the injection of fentanyl at the dose of 150 μg/kg or 300 μg/kg (infused over 1 min) in every animal. Note that both doses produced a similar effect, whereas 13 animals displayed an immediate increase in Prs by more than 20%. B2 and B3: change in Prs in keeping with the dose of ketamine administered in each animal (initial dose, B2) and total dose (B3). There was no relationship between the doses of ketamine and the magnitude of the response to fentanyl. Prs, trans-thoraco-pulmonary pressure. *P < 0.05, when compared to control.

Protocol

Tidal volume and breathing frequency were set at 7.5 mL/kg (3.87 ± 0.23 mL) and 70 breaths/min, respectively, and this breathing pattern was maintained constant throughout the entire period of the study. After a period of at least 15 min of stability under mechanical ventilation, fentanyl was infused at 150 or 300 μg/kg over 1 min. Since, as described in the results, animals that displayed two different patterns of response, saline or dexmedetomidine, were administered at two different timings, depending on the type of response, i.e., a rapid and abrupt increase in Prs versus a more progressive rise in Prs over time. In the former, dexmedetomidine or saline was administered 5 min after fentanyl, and the changes in Crs, Rrs, τrs, and V̇o₂ were determined before and throughout the 30 min-period following fentanyl infusion (see Data Analysis for the timing of comparison), whereas arterial blood (ABG) was sampled before (T₀) and after fentanyl administration, i.e., 3 min after fentanyl (T₃; before saline or dexmedetomidine), and 30 min after fentanyl injection (T₃₀). In the rats that displayed a progressive decrease in Crs, dexmedetomidine was administered 30 min after fentanyl injection. In these animals, ABG was sampled with the same timing as earlier, but an additional ABG was sampled 5 min after dexmedetomidine.

Data Analysis

Our primary outcomes were the changes in the compliance of the respiratory system (Crs), in V̇o₂, arterial partial pressure of oxygen ($\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$), and arterial partial pressure of carbon dioxide ($\text{P}{\text{a}}_{\text{C}\text{O}}{}_{{}_2}$). Data were presented as means ± standard deviation in the text for clarity; however, individual data were shown in figures along with a box and whiskers representation (median, minimum, and maximum values). Crs were compared before (T₀) and after fentanyl (T₃, T₉, and T₃₀) for the animals displaying a very abrupt rise in Crs and at T₀, T₉, T₃₀, and T₃₅ for the other group, using a Friedman test (paired two-tailed test). Whenever a significant difference was found, a Wilcoxon (W) matched-pairs signed-rank test was used to compare any of these specific time points. Comparison between groups was done using a Kruskal–Wallis test for multiple comparisons and a Mann–Whitney (MW) U test for two-group comparisons. Statistical analyses were performed using GraphPad Prism 5 (Graphpad Software, La Jolla, CA).

RESULTS

Fentanyl injection, whether at the dose of 150 (n = 10) or 300 (n = 13) μg/kg, produced an immediate suppression of any residual spontaneous inspiratory activity, absence of negative inflection in the pressure signal throughout the period of study. All animals were therefore considered apneic, while mechanically ventilated. Since the effects on Prs were the same regardless of the doses of fentanyl (Fig. 5), the response to fentanyl was combined for the rest of the analysis. Finally, the doses of ketamine (induction or total doses) that were administered, were unrelated to the effects of fentanyl on Prs (Fig. 5). Fentanyl always produced a decrease in Crs, but with two clearly different temporal profiles as shown in Figs. 6 and 7.

Figure 6.

Example of the effects of an injection of fentanyl (300 µg/kg over 1 min) on respiratory flow, tidal volume, and Prs in one mechanically ventilated rat (volume control and constant flow). One minute after fentanyl injection, Prs increased very rapidly along with a short-lasting phase of rhythmic contractions of most skeletal muscles that were transmitted to the flow and pressure signals. This phase, which was observed in 13 out of 23 rats that were studied, was of very short duration and was replaced by persistent tetanic/tonic contractions of skeletal muscles mimicking those observed in unsedated animals (Fig. 4). This effect was associated with reduction in the respiratory system compliance (Crs) that persisted for the entire duration of the study. A, B, and C are magnification of selected portions of the signals. Note 1) generation of normal breaths by the ventilator was impeded during the short-lasting initial phase of rhythmic contractions. This portion of the recording was always excluded from the analysis; 2) following this startling response, inspiratory flow and volume returned to baseline and Prs remained elevated, along with a persistent contraction of a large mass of skeletal muscles (see text for more details) with no spontaneous inspiratory activity. Crs, compliance of the respiratory system; Prs, trans-thoraco-pulmonary pressure.

Figure 7.

Example of the effects of an injection of fentanyl (300 µg/kg over 1 min) on respiratory flow, tidal volume, and Prs in one mechanically ventilated rat (volume control and constant flow), which in contrast to the rat shown in Fig. 6, displayed a progressive increase in Prs (and decrease in Crs). This effect was always associated with an increase in V̇o₂ reflecting the metabolic consequences of the skeletal muscle contractions. This type of response (absence of the initial and short-lasting periods of volleys of phasic contractions) was displayed by 10 out of the 23 rats studied. Crs, compliance of the respiratory system; Prs, trans-thoraco-pulmonary pressure; V̇o_2, oxygen consumption.

Rats Displaying an Immediate Increase in Prs

In 13 out of 23 rats (56% of the animals), a very abrupt increase in Prs occurred within 90 s following the end of fentanyl injection (Figs. 5 and 6). This startling effect was associated with phasic and rhythmic contractions of most skeletal muscles. This abrupt initial response subsided within 40 s and was replaced by a continuous tonic/tetanic contracture of most skeletal muscle groups. These sustained tetanic contractions produced an adduction of the forelegs, a flexion of the hind legs, and a persistent contraction of the abdominal and back muscles (opisthotonos with extension of the tail), akin to the effects observed in the nonsedated animals (Fig. 4). When the rhythmic muscular activity ceased, Crs was decreased by half, dropping from 0.51 ± 0.11 mL/cmH₂O (R² values of the P-V relationship averaged 0.989 ± 0.006) to 0.26 ± 0.06 mL/cmH₂O (R² values of the P-V relationship averaged 0.969 ± 0.018). Crs averaged 0.36 ± 0.08 mL/cmH₂O (R² values of the P-V relationship averaged 0.992 ± 0.007, median = 0.995) at T₃, as illustrated in Figs. 8 and 9 and shown for all data in Fig. 10. The changes in Crs were significantly lower than baseline at T₃ when compared with T₀ (Fig. 10, P values are given in the figures). In this group of 13 rats, six received the saline, whereas seven animals received dexmedetomidine under the same volume, 5 min after the end of fentanyl administration.

Figure 8.

Examples of the changes in respiratory flow, volume (VT), and Prs produced by fentanyl 300 μg/kg in one mechanically ventilated rat that presented a startling response. Baseline data (before fentanyl) are in black, whereas the data recorded in steady state conditions ∼3 min after fentanyl administration are shown in red. A is the period of inspiration, B is expiration. Fentanyl resulted in a significant rise in driving pressure generated by the ventilator to maintain the same flow and volume (A). B: the left figure shows that Crs values, computed as the slope of the pressure-volume relationship (established from data shown in A) decreased after fentanyl. The same effects can be seen when data were analyzed using multiple regression method (middle and right). Averaged data are given in the text. C: typical flow expiratory patterns in baseline and after fentanyl administration (∼3 min after fentanyl infusion) in the same animal. D: the passive time constant of the respiratory system (τrs) is shown for each condition, as the slope of the linear part of the expiratory flow-volume curve. V₀ (see also Fig. 3) represented as a negative volume, corresponds to the volume actively maintained above the passive position of the relaxation of the respiratory system, i.e., passive functional residual capacity (FRC_passive). C: note that 1) following fentanyl, the time constant of the respiratory system is much shorter, reflecting the decrease in Crs; 2) peak expiratory flows after fentanyl were more negative than in baseline condition, whereas end expiratory flow were much closer to zero, reflecting a reduction in the volume maintained above FRC_passive. These data were obtained from the group of 13 rats (out of 23) that presented an immediate response as illustrated in Fig. 6. Prs, trans-thoraco-pulmonary pressure.

Figure 9.

Examples of recordings in two different ketamine-sedated mechanically ventilated rats showing the changes in respiratory flow, tidal volume, and Prs produced by 300 μg/kg fentanyl followed by saline or dexmedetomidine. The respiratory system compliance decreased (increased Prs for same flow and volume) in the untreated animal; dexmedetomidine (3 μg/kg), administered 5 min after fentanyl injection, decreased Prs, and rescued the respiratory compliance and V̇o₂. Note that in the rat receiving saline, abrupt and sporadic changes in the pressure signal (*) were observed, which as shown in A and A' can be accounted for by brief and sporadic active expiration. This phenomenon which had no regularity was not observed after dexmedetomidine. Crs, compliance of the respiratory system; Prs, trans-thoraco-pulmonary pressure; V̇o_2, oxygen consumption.

Figure 10.

Box and whiskers representation (median, minimum, and maximum values) with individual data of the change in Crs, V̇o₂, and arterial blood pressure at different timings before (T₀) and after fentanyl infusion (T₃) in animals receiving saline or dexmedetomidine (at T₉ and T₃₀). $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$ and $\text{P}{\text{a}}_{\text{C}\text{O}}{}_{{}_2}$ are also shown before (T₀) and after fentanyl (T₃) as well as after saline or dexmedetomidine (T₃₀). These data were obtained from the 13 rats that displayed an immediate response (decrease in Crs, see Fig. 6): seven of these animals received saline, and six animals received dexmedetomidine at T₅. In the nontreated animals, there was a continuous reduction in V̇o₂ and a significant hypoxemia, which was rescued by dexmedetomidine. A Friedman (F) test was used to compare the paired data for a given condition (saline or dexmedetomidine), P values are given. A Wilcoxon (W) test was used to compare matched pairs of any specific time points vs. baseline (*P < 0.05, **P < 0.001). Whenever other specific time points were compared, P values are given. A Mann–Whitney (MW) test was used to compare data from the two groups ($P < 0.05, $$P < 0.01). Finally, note that the changes in Crs were completely dissociated from the change in blood pressure. Crs, compliance of the respiratory system; $\text{P}{\text{a}}_{\text{C}\text{O}}{}_{{}_2}$, arterial partial pressure of carbon dioxide; $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$, arterial partial pressure of oxygen; V̇o₂, oxygen consumption.

Saline Group

Crs continued to subside until the end of the 30-min period of the study, reaching 0.31 ± 0.03 mL/cmH₂O (−41% from baseline) at T₃₀ (Figs. 9 and 10). This decrease in compliance was significantly lower at T₉ and at T₃₀ than at T₀ (Fig. 10). However, there was a spontaneous return of Rrs to baseline values at T₃₀ (0.348 ± 0.333 cmH₂O/mL/s, not different from baseline). This led to a passive expiratory time constant that was shorter than in baseline condition (Δτrs = −160 ± 100 ms or −41 ± 19%), along with a reduction in V₀ (see Eq. 4) as shown in Fig. 8 due to the persistent reduction in Crs. These mechanical changes were associated with the persistence of tonic contractions of most skeletal muscles, that lasted during the entire period of the study (abdominal muscles, abduction of the forelimbs, and extension of the hindlimbs). This muscular activity was also associated with a significant increase in V̇o₂, which rose from 8.48 ± 4.31 (baseline) to 11.29 ± 2.57 mL/min at 30 min (Figs. 9 and 10). V̇o₂ was significantly higher than baseline at T₃₀ (P values given in Fig. 10). This increase in metabolism was associated to a significant decrease in $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$ at T₃₀ (Fig. 10).

Effects of the α₂ Agonist Dexmedetomidine

In marked contrast to the effects of saline (Figs. 9 and 10), the administration of dexmedetomidine immediately interrupted the decline in Crs, which increased toward the baseline values reaching 0.46 ± 0.02 mL/cmH₂O at 30 min. Crs values were not different at T₉ and T₃₀ than at baseline (T₀) and were significantly different from the saline group at T₃₀ (−23 ± 14% vs. −53 ± 5% of baseline, P values are given in Fig. 10). Of note, in four out of seven treated rats, dexmedetomidine rescued Crs to more than 80% of the baseline values. These changes in Crs were associated with a reduction in Rrs at T₃₀, which returned even below baseline levels (0.258 ± 0.142 cmH₂O/mL/s). The decrease in Rrs offset the partial restoration of Crs leading to a time constant, which remains lower than baseline (−113 ± 80 ms or –32 ± 14%). In all dexmedetomidine-treated animals, V̇o₂ decreased below baseline (Figs. 9 and 10), instead of rising, reaching 6.4 ± 1.4 mL/min at T₃₀ (vs. 11.3 ± 2.6 mL/min in the saline group). The difference between the two groups was significant at T₉ and T₃₀. This was associated with a restoration of arterial blood gases at T₃₀ as shown in Fig. 10.

Additional Observations

As shown in Fig. 6 or 9, we did not observe any spontaneous depression in the pressure signal reflecting an attempt from the animals to initiate a spontaneous inspiratory activity following fentanyl administration, nor did we observe a systematic asynchrony on the pressure signal. However, as shown on Fig. 9, rare and sporadic “overshoot” on the pressure signal were observed, which could be accounted for by a brief active expiratory effort typically for a few breaths. This phenomenon, which had no regular pattern, was always associated with general tonic contractions of a large mass of muscle and disappeared after dexmedetomidine.

As illustrated in Fig. 8, peak expiratory flow after fentanyl was much more negative than in baseline condition (likely due to the higher recoil pressure of the respiratory system), whereas end-expiratory flow was much closer to zero after fentanyl than before, reflecting a shortening of the expiratory time constant.

Finally, blood pressure changes produced by fentanyl or dexmedetomidine were completely dissociated from the change in Crs (Fig. 10). Of note, dexmedetomidine produced a transient increase in blood pressure that was significantly higher at T₉ than at T₃ (Fig. 10), followed by a moderate reduction in mean arterial pressure.

Rats Displaying a Progressive Decrease in Crs

As illustrated in Figs. 7 and 11, in the 10 remaining rats that did not display an abrupt increase in Prs after fentanyl (Prs changed by less than 20% within the first 2 min), a progressive increase of Prs was observed over time, reflecting a progressive decrease in Crs throughout the 30-min period of the study (Crs significantly decreased from 0.40 ± 0.06 mL/cmH₂O at T₀ to 0.23 ± 0.07 mL/cmH₂O at T₃₀, see Fig. 12 for details and P values). These mechanical changes were associated with a significant increase in V̇o₂ (from 7.9 ± 2.1 to 9.7 ± 2.3 mL/min, Figs. 7, 11, and 12) associated to a significant decrease in $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$ and increase in $\text{P}{\text{a}}_{\text{C}\text{O}}{}_{{}_2}$ (Fig. 12). Administration of dexmedetomidine, 30 min after fentanyl injection, produced a rapid, i.e., occurring during the period of injection, and significant decrease in V̇o₂ and increase in Crs (Figs. 11 and 12) in 9 out of the 10 animals (see Fig. 12 for statistical analysis). Arterial oxygen saturation ($\text{S}{\text{a}}_{{\text{O}}_2}$) and $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$ after dexmedetomidine were also significantly higher than at T₃₀, whereas $\text{P}{\text{a}}_{\text{C}\text{O}}{}_{{}_2}$ significantly decreased.

Figure 11.

Example of the effects of an administration of dexmedetomidine (3 μg/kg) in a rat that displayed a progressive increase in Prs (and decrease in Crs) after an injection of fentanyl (300 μg/kg over 1 min). Note that dexmedetomidine increased Crs and V̇o₂. Like in Fig. 10, this rat presented sporadic episodes of active expiration, visible on the pressure signal (*), which were abolished after dexmedetomidine. Crs, compliance of the respiratory system; Prs, trans-thoraco-pulmonary pressure; V̇o₂, oxygen consumption.

Figure 12.

Box and whiskers representation (median, minimum, and maximum values) with individual data of the change in Crs, V̇o₂, $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$, $\text{P}{\text{a}}_{\text{C}\text{O}}{}_{{}_2}$, and $\text{S}{\text{a}}_{{\text{O}}_2}$ at different timings before (T₀), after fentanyl infusion (T₉ and T₃₀), and after dexmedetomidine (T₃₅). These data were obtained in the group of 10 remaining rats (out of 23) that presented a progressive decrease in Crs, as illustrated in Fig. 7. Note that for blood gas analysis, data were only compared using the 8 rats in which blood could be collected at every time point (the 2 animals that were missing data at T₉, T₃₀, or T₃₅ were excluded from the analysis). Crs, compliance of the respiratory system; $\text{P}{\text{a}}_{\text{C}\text{O}}{}_{{}_2}$, arterial partial pressure of carbon dioxide; $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$, arterial partial pressure of oxygen; $\text{S}{\text{a}}_{{\text{O}}_2}$, arterial oxygen saturation; V̇o₂, oxygen consumption. *P < 0.05, **P < 0.01.

DISCUSSION

We found that fentanyl-induced muscle contracture/rigidity has two major respiratory consequences: 1) a prolonged decrease in the compliance of the respiratory system and 2) an increase in metabolism, resulting from the contractions of a large mass of skeletal muscles, which persisted for at least 30 min. As developed in the following paragraphs, the impact of the mechanical changes produced by a fentanyl overdose on the capacity to maintain alveolar ventilation along with the alteration in metabolism—affecting in turn the alveolar gas equation—could be, by themselves, life-threatening and contribute to the development of critical levels of hypoxemia.

Opioid Intoxication-Induced Breathing Depression and Alterations in Respiratory Mechanics

The mechanisms of opioid-induced breathing depression are complex. First, large doses of opioid agonists can produce a direct depression of the activity of medullary respiratory neurons “equipped” with μ-opioid receptors in the preBötzinger complex (38–41) as well as in pontine structures (42, 43). The hyperpolarization of these respiratory neurons, in the presence of an opioid agonist (44), inhibits the cyclic generation of breaths. Second, opioid-induced sedation develops at lower doses than those directly inhibiting medullary respiratory neurons and can produce a depression in breathing (45) in keeping with the level of metabolism and arousal. Such a depression occurs via the inhibition of the activity in the reticular formation, including the periaqueductal gray matter, along with an alteration in the level of vigilance and arousal (46, 47). Finally, opioid-induced muscle rigidity can produce a unique, detrimental, and potentially lethal effect (5, 7–10, 48). We found that muscle rigidity induced by high-dose fentanyl led to a long-lasting decrease in Crs over the entire 30-min period of the study. Of note, baseline values of Crs were similar to those already reported in anesthetized rats, in keeping with body weight (49, 50). The decrease in Crs was temporally associated with tonic contractions of many skeletal muscles including tetanic contractures of abdominal and lumbar muscles, adduction of the upper limbs, and flexion of the lower limbs. However, in major contrast to what was produced in the nonsedated animals in our previous studies (18, 19), sedation prevented the recovery of spontaneous inspiratory activity. We could never identify any spontaneous depression on the pressure signal, nor did we observe regular asynchrony or spontaneous expiratory activity on the pressure signals, except in the animals presenting an initial startling response (rapidly reversible rhythmic volleys of rhythmic contractions as illustrated in Fig. 6). We did, however, observe over time, sporadic jerking episodes affecting one or two breaths that would create an artifact on the pressure signal during expiration (Figs. 9 and 11).

Following the description of Hamilton and Cullen in 1953 (3), opioid-induced muscle rigidity is now well established in humans. It has been observed during postoperative periods following the use of fentanyl, for up to several hours (48). Of interest, low-doses intravenous fentanyl (1–2 μg/kg), which do not necessarily produce a central apnea, can generate severe truncal and abdominal muscle rigidity. Similarly, Jerussi et al. (21) have reported that muscular rigidity in rats can be induced by the intravenous administration at 35 μg/kg fentanyl, which is nonfatal in these species. The potential consequences of the changes reported in the present study should, however, be much more prominent during an opioid overdose in an unresponsive subject with depressed ventilation (4), in whom a bolus injection of opiates and a high-speed injection can precipitate the onset of muscle rigidity. Fentanyl-induced chest rigidity has been shown to make patients difficult, or sometimes almost impossible, to mechanically ventilate (51). The effect can be so dramatic that it can require neuromuscular blockade, intubation, and/or naloxone to save the patients. For instance, a study performed on 30 subjects showed that remifentanil (1 μg/kg) provokes very difficult mask ventilation in 70% of patients within minutes following the injection (51). However, no data reporting the change in compliance versus upper-airway resistance were available in these patients (51).

Opioid-induced glottic closure and increased resistance in the upper airways have also long been recognized (52–54). The contribution of increased resistance of the upper airways typically due to a partial and intermittent glottic closure (53) was not evaluated in the present study. We did however observe a modest and spontaneously reversible increase in lower Rrs following fentanyl. This effect may have resulted from a possible bronchoconstrictive effect of opioid, a phenomenon suspected in patients with asthma (for review, see Ref. 55). However, such a proposition appears to be contradicted by data showing that the stimulation of opioid receptors decreases reflex (56) or nonadrenergic noncholinergic bronchoconstriction (57). Alternative explanations could be offered since Rrs includes airway resistance along with the lung and chest wall tissular resistances; an increase in Rrs may have, therefore, reflected an increase in tissular resistance, possibly of the chest wall, illustrating an inverse proportionality between the change in chest wall resistance and compliance. Finally, as discussed in the next paragraph, the reduction in functional residual capacity (FRC) associated with the faster time constant of the respiratory system could have resulted in an increased resistance simply due to the lower lung volumes at which measurements were made (Fig. 8).

One can speculate that whenever truncal and abdominal muscle rigidity develops (5) on top of an already severely depressed breathing drive, a potentially lethal reduction in tidal volume can certainly occur. Specific populations of patients may be at risk when exposed to a decreased chest wall compliance (such as in patients with obesity or in patients with obstructive or restrictive limitations), who can fail to maintain eupneic breathing in the presence of an increased elastic load to breathe.

Crs and FRC following Fentanyl

In small mammals, like in the rat, a high chest wall compliance (58–60) requires the functional residual capacity (FRC) to be actively maintained (FRC_active) above the passive position of relaxation of the respiratory system (FRC_passive). A spontaneous high frequency of breathing, the narrowing of upper airways during expiration and the post inspiratory activity on the diaphragm during the first part of expiration create an intrinsic PEEP during spontaneous breathing and maintain FRC_active above FRC_passive (for review, see Refs. 50, 60).

“Actively” maintaining FRC is critical to prevent atelectasis and hypoxemia from developing as lung volumes decrease in small mammals. In our model, since animals were tracheostomized and sedated, some of the mechanisms maintaining FRC_active (no upper airway narrowing) cannot operate. It is therefore mainly through the breathing frequency that FRC_active could be maintained in our mechanically ventilated rats. As shown in Fig. 8, in baseline condition, expiratory flow never reached zero before a new inspiration starts, reflecting the presence of a significant intrinsic PEEP. A decrease in the time constant by ∼160 ms after fentanyl (while the expiratory time was less than 400 ms) could have certainly resulted in a reduction in FRC_active, as shown in Fig. 8, wherein end-expiratory flow was closer to zero, and V₀ (the extrapolation of the flow-volume linear relationship during expiration at zero flow) was significantly lower after fentanyl. This reduction in FRC should certainly be considered as an additional factor of hypoxemia (lower lung volume, as discussed in the last paragraph of this section) and could have also created a viscous circle, leading to a further decrease in Crs.

Muscle Rigidity-Induced Hypermetabolism: The Alveolar Gas Equations during Opioid Overdose

How much do we expect the metabolism to increase because of opioid-induced muscle rigidity? We found in our sedated rats that V̇o₂ can increase by more than 50% during opioid-induced muscle rigidity. This can be explained by the large number of skeletal muscles involved, including the back (opisthotonos) and the trunk muscles. Adduction and flexion of the forelegs and hindlimbs were also consistently present. Studies performed on human subjects receiving intravenously opioids also showed that a very large muscle mass is involved for long periods of time (48).

As minute ventilation was maintained constant, the decrease in $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$ that we observed after fentanyl can therefore be in part accounted for by the increase in V̇o₂. Indeed, in keeping with Eq. 5, alveolar partial pressure in O₂ ($\text{P}{\text{A}}_{{\text{O}}_2}$) is dictated by the ratio between oxygen consumption (uptake; V̇o₂) and alveolar ventilation (V̇A; 31, 32).

Figure 13 illustrates the impact of the level of metabolism on $\text{P}{\text{A}}_{{\text{O}}_2}$, when, in real live situation, breathing generation is also impaired and hypoventilation occurs. For instance, in an adult with a resting V̇o₂ of 300 mL/min and a baseline alveolar ventilation of 5 L/min, a decrease in alveolar ventilation to 2.5 L/min (decrease by half of alveolar ventilation) will decrease $\text{P}{\text{A}}_{{\text{O}}_2}$ from ∼100 to ∼50 mmHg. If the same changes are associated to a doubling in V̇o₂, $\text{P}{\text{A}}_{{\text{O}}_2}$ will reach zero. In other words, the level of hypermetabolism can account for a significant part of the hypoxemia observed in our animals. The involvement of other mechanisms, e.g., pulmonary edema, that could potentially affect V/Q matching or O₂ diffusion is reviewed in Limitations of the Experimental Model and Confounding Factors.

Figure 13.

Representation of the changes in $\text{P}{\text{A}}_{{\text{O}}_2}$ as a function of alveolar ventilation (V̇A) in a resting adult human. Note that for a given decrease in alveolar ventilation, here from 5 to 3.5 L/min, the resulting $\text{P}{\text{A}}_{{\text{O}}_2}$ is critically dependent of the level of V̇o₂. Severe hypoxemia can develop, even as a result of a moderate increase in metabolism, whenever an additional depression in ventilation would be present. $\text{P}{\text{A}}_{{\text{O}}_2}$, alveolar partial pressure of oxygen; V̇o₂, oxygen consumption.

Effect of Dexmedetomidine

Neurons in the locus coeruleus appear to be critical to produce opioid-induced muscle contractions in mammals [11–16; and possibly for opioid-induced automatic locomotion in mice (61)]. Indeed, this small region of the rostral medulla is involved in the control of muscle contractions, reducing muscle tone via an inhibiting adrenergic pathway. This effect is suppressed by fentanyl, producing in turn an increased muscle rigidity. A balance between the stimulation of μ-, δ-, and κ-receptors has been proposed to be a determinant of the production of opioid-induced muscle rigidity (62). α-2 agonists have been shown to restore a normal pathway in opioid intoxicated models and to inhibit muscle rigidity (15, 17, 21). The effects on breathing of reducing muscle rigidity are not well characterized, some studies nevertheless suggested a potential beneficial effect. For instance, Alfentanyl at the dose of 500 μg/kg subcutaneously produces a decrease in inspiratory and an increase in expiratory diaphragmatic electromyogram (EMG) activity (63). Although these changes in diaphragm function were not accompanied by significant respiratory depression, the α-2 agonist dexmedetomidine at the dose of 30 μg/kg significantly attenuated alfentanyl-induced increase in expiratory muscle activity (63). In human studies, Park et al. (51) have found that the incidence of difficult mask ventilation, after opioid, i.e., requiring higher driving pressure to deliver a given tidal volume, was significantly lower in dexmedetomidine treated patients than in the control group. i.e., 36.7% versus 70.0%.

Limitations of the Experimental Model and Confounding Factors

Compared with humans, the doses of fentanyl used in rats to produce a complete apnea are very high when expressed/kg—and incidentally are even higher in mice—i.e., approximately, about hundreds of microgram per kilogram in rats and tens of milligrams per kilogram in mice (64). These doses are much higher than those used in humans for analgesia (65) or for procedural sedation (66). They nevertheless correspond to the doses of fentanyl used in certain protocols like for induction of general anesthesia in cardiac surgery (67), or most realistically during an opioid overdose. Experiments using opioids in rodents must therefore be analyzed in keeping with the physiological effects that are produced, rather than the dose/kg per se (68), since doses of opioid in small mammals do not follow the expected allometric relations among species (34, 35).

In contrast, the dose of dexmedetomidine appears to be compatible with low doses used in humans. Indeed, 1 μg/kg dexmedetomidine in humans is able to overcome the effects of difficult ventilation after opioid (51). This dose of dexmedetomidine corresponds to the doses that we used in our rat model, in keeping with the equivalence between doses in rats and those used in humans [7–10 times higher doses/kg in rats than in humans (34, 35)]. These doses are therefore clinically relevant as they can prevent agitation (69, 70) without producing any depression in breathing in humans. Of note, doses up to 30 μg/kg of dexmedetomidine do not seem to produce any hypoventilation in rats (63).

One important limitation of our study certainly resides in the fact that the animals were sedated. Ketamine was used since it has been shown to be one of few sedative agents preserving opioid-induced rigidity (7). In that study, Lui et al. found that intravenous administration of fentanyl (25, 50, or 100 μg/kg) caused muscle rigidity in ketamine, but not in thiopental, anesthetized animals. In addition, hypoxemia was present in both spontaneously ventilating and ketamine-anesthetized rats, but not in thiopental-anesthetized animals. The authors concluded that the use of ketamine in animals provide an alternative approach in the evaluation of fentanyl-induced muscle rigidity, whenever sedation is needed. Ketamine, at anesthetic dosing, can produce muscle contractions; this effect is not constant (71–73; observed in 2%–16% of animals). We found no evidence that fentanyl-induced muscle rigidity could have been affected (sensitized) by ketamine: there was no relationship between the doses of ketamine received by the animals and the intensity of the change in Prs (Fig. 5). Perhaps more importantly, the muscular response that we observed in our sedated rats perfectly mimicked the time course and the clinical presentation of what we observed in unsedated rats (Fig. 5).

In an adult human, difference in the partition between the chest wall and lung compliance, when compared with other mammals, must be recognized (74). The compliance of the lungs (CL) is similar to the compliance of the chest wall (Cw) in adult humans, in contrast to newborns (75) and infants (74, 76) as well in nonhuman mammals (58–60), all displaying a higher Cw than CL (77, 78). Since 1/Crs = 1/CL+ 1/Cw (79), a decrease by half of Cw for instance will have a much larger impact on Crs (and thus on the elastic load to breathe) in an adult human than in any nonprimate mammals, in turn requiring a higher drive to breathe to maintain the same tidal volume.

Finally, other factors can alter the respiratory mechanics following an opioid overdose. Opioid-induced noncardiac pulmonary edema may certainly decrease CL and therefore Crs (80) producing an acute onset of hypoxic respiratory failure (81) due to shunting and ventilation perfusion mismatch. The rapid effect of dexmedetomidine on Crs, V̇o₂, and $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$ supports the view, however, that a large part of the changes in Crs are due to a decrease in Cw. Nerveless, the fact that dexmedetomidine did not produce a full recovery of Crs may indicate that low FRC-induced microatelectasis or even pulmonary edema can develop over time and contribute to the persistence of a reduced Crs. Of note, we have not reported direct measurements of Cw in the present paper, although Cw was monitored in some of our animals, using esophageal pressure. In our hands, esophageal pressure was very variable, making it difficult to determine, with enough confidence, the changes in lung versus chest wall compliance produced by fentanyl.

Relevance Of The Present Results to Fentanyl Overdose In Humans

Torralva and Janowsky (82) have hypothesized that the significant increase in the number of deaths by fentanyl (vs. morphine for instance) may be accounted for by adrenergic and cholinergic receptor-mediated mechanical failure of the respiratory system due to the rapid development of rigidity, rather than or in addition to a primary centrally mediated ventilatory depression. Their rationale was that a lethal outcome is not as typical with high doses of morphine or its prodrug, heroin, which only cause mild rigidity in abdominal muscles, in contrast to fentanyl. Burns et al. (5) have also proposed that a sudden onset of chest wall rigidity may represent an underestimated mechanism of mortality from illicit intravenous fentanyl use. The present study certainly supports the view that the effects of fentanyl on the respiratory mechanics can contribute to fentanyl lethality by very rapidly affecting the load to breathe with a significant impact on the metabolic response and therefore on blood gas hemostasis.

The respiratory control system is normally able to adjust its drive to the change in elastic load (decreased compliance) via vagal and thoracic afferent signals (83), as well as via a cortical response. In the presence of a severe depression by opioids, any additional increase in the mechanical load to breathe (84) can certainly be challenging for the respiratory control system preventing the generation appropriate tidal volume. Also, minute (alveolar) ventilation is constantly adjusted by the respiratory control system to changes in metabolic rate, in turn keeping blood gas homeostasis (31, 32). However, an acute increase in metabolic load could further reduce alveolar Po₂ if a victim already hypoventilating. The effects observed in the present study could, in and of themselves, certainly contribute to fentanyl lethality (5). These adverse effects appear to be mitigated by dexmedetomidine, with a significant improvement in the level of oxygenation. Studies in nonsedated animals and in larger mammals must confirm the clinical significance of the present findings.

DATA AVAILABILITY

Data will be made available upon reasonable request.

GRANTS

This work was supported by the National Institute of Health, Grant No. 1R61HL156248-01.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

P.H. conceived and designed research; P.H. and N.T. performed experiments; P.H. and N.T. analyzed data; P.H. interpreted results of experiments; P.H. prepared figures; P.H. drafted manuscript; P.H. and N.T. edited and revised manuscript; P.H. and N.T. approved final version of manuscript.