Fentanyl Overdose: Temporal Effects and Prognostic Factors in SKH1 Mice

Abstract

Fentanyl exposure and overdose are growing concerns in public health and occupational safety. This study aimed to establish parameters of fentanyl lethality in SKH1 mice for future overdose research. Lethality was determined using the up-down procedure, with subjects monitored post-administration using pulse oximetry (5 minutes) then whole-body plethysmography (40 minutes). Following determination of subcutaneous dose-response, [18F]Fluorodeoxyglucose Positron Emission Tomography (¹⁸F-FDG PET) was performed after LD10 fentanyl at 40 minutes, 6 hours, 24 hours, or 7 days post-dose. LD10 and LD50 were observed to be 110 and 135 mg/kg, respectively, and consistent with four-parameter logistic fit values of 111.2 and 134.6 mg/kg (r²=0.9996). Overdose (LD10 or greater) yielded three distinct cardiovascular groups: survival, non-survival with SpO2 minimum ≥37%, and non-survival with SpO2 <37%. Breaths per minute, minute volume, and inspiratory quotient were significantly different between surviving and non-surviving animals for up to 40 minutes post-injection. ¹⁸F-FDG PET revealed decreased glucose uptake in heart, lungs, and brain for up to 24 hours. These findings provide critical insights into fentanyl lethality in SKH1 mice, including non-invasive respiratory effects and organ specific impacts that are invaluable for future translational studies investigating the temporal effects of fentanyl overdose.

Simple English Summary

Fentanyl exposure and overdose are growing concerns in public health and occupational safety. This study aimed to establish parameters of fentanyl lethality in SKH1 mice for future overdose research. After a single subcutaneous injection of fentanyl, subjects were monitored via pulse oximetry for 5 minutes, respiration for 40 minutes, then imaged for glucose uptake using Positron Emission Tomography (PET). Pulse oximetry revealed a cardiac phenomenon determining acute lethality, respiratory data exhibited significant differences associated with survival, and PET showed decreased glucose uptake for up to 24 hours. These non-invasive, organ-specific findings are critical for future temporal studies on fentanyl overdose.

Introduction

Fentanyl was first synthesized in 1960 and implemented as an anesthetic in 1963¹; since its inception, it has seen an explosion in popularity ranging from surgery recovery to cancer pain management¹ and now worldwide abuse for its analgesic effects^2,3. While exceedingly beneficial in the clinical arena, its high potency, rapid onset, and duration of action contribute to the public health risk of overdose deaths^4,5.

Fentanyl has surpassed all other opioids combined in overdose deaths^6,7 and was the leading cause of deaths in individuals aged 18–45 in the United States in 2020 and 2021⁷. Opioid-induced respiratory depression (OIRD) causes brain hypoxia that contributes significantly to the lethality associated with fentanyl^8,9. The region of the brain responsible for breathing, the medulla oblongata, has a high density of μ-opioid receptors¹⁰; as fentanyl saturates receptors, respiration decreases, resulting in systemic hypoxia, particularly in the brain¹¹. OIRD can typically be reversed with μ-receptor antagonists such as naloxone, if administered quickly. However, fentanyl overdose can be resistant to naloxone, unlike classical opioids (e.g., morphine, codeine, and heroin), potentially due to a largely uncharacterized response known as Wooden Chest Syndrome (WCS)¹². Broadly consisting of chest muscle and airway rigidity combined with laryngospasm¹³, WCS appears to be unique to fentanyl and its analogs. Given the number of attributable mortalities annually combined with the limited understanding of overdose physiology, basic research with animal models is needed to explore the systemic impacts of high doses of fentanyl.

This study explores the overdose lethality of subcutaneously-administered fentanyl in SKH1 mice, which are a hairless yet immunocompetent, outbred strain ideal for translational safety and efficacy studies^14–19. Subcutaneous exposure and overdose, as the result of missed intravenous injection or as a surrogate of dermal exposure, are an increasing concern in public health and occupational safety^20–22. The subcutaneous median lethal dose (LD50) of fentanyl, 62 mg/kg, was measured in 1964 in Swiss-Webster mice²³ and is commonly referenced in Safety Data Sheet (SDS) documentation (though later publications reported 7.57²⁴ and 28.5²⁵ in the same strain) but has not been determined in SKH1 mice. Beyond determining the lethality of fentanyl in SKH1 mice, the present study displays physiological outcomes of bolus fentanyl overdose at acute, intermediary, and longer-term time points and identifies prognostic factors that determine survival.

In the acute period of intoxication (up to 5 minutes post-administration), changes in heart rate (HR) and blood oxygen saturation (SpO2) were measured and their associations with mortality were identified. In the intermediary period (5–40 minutes post-administration), the same associations were measured in animals who survived past the first 5 minutes via plethysmography. In the long-term period, at 40 minutes, 6 hours, 24 hours, and 7 days post-administration, positron emission tomography with [18F]Fluorodeoxyglucose (¹⁸F-FDG PET) was used to visualize the distribution of changes in glucose uptake in response to fentanyl, and these effects were verified in individual organs ex vivo via residual radioactivity.

This study provides valuable insights into the lethality and physiological effects of fentanyl overdose. For all opioids, roughly 11–14% of overdoses are fatal²⁶; statistics regarding fentanyl-specific overdose survival are currently unavailable. Identification of prognostic factors associated with survival from this study could aid in the development of more effective treatments for fentanyl overdose. Additionally, the development of mouse models is crucial in understanding the systemic impacts of high doses of fentanyl and for developing potential therapies. Previous fentanyl research in animal models primarily addresses addictive properties^27–29 and protective treatments^30,31 at doses that are therapeutically relevant, but seldom addresses the temporal outcomes or organ-wide effects of overdose. The findings of this study have significant implications for public health and occupational safety and highlight the need for continued research in this area.

Materials and Methods

The study was conducted in accordance with the Basic & Clinical Pharmacology & Toxicology policy for experimental and clinical studies³². All procedures were evaluated and approved by the Institutional Animal Care and Use Committee at West Virginia University (WVU IACUC) and USAMRMC ACURO.

Animals

Adult SKH1 Elite mice (strain code 477, Charles River, Wilmington, MA; n = 46 for all procedures except PET, where n = 11), aged 11–14 weeks (23–36 grams), were used in these experiments, and were group housed (5 per cage) in a vivarium at 40–60% humidity and 20–24°C with a 12-hour light/dark cycle. All experiments were conducted during the light portion of the cycle in a surgical suite. Animals had ad libitum access to Envigo Teklad 2918 chow and water. An acclimation period of at least three days was provided for all animals. Individual subjects were randomly selected during experiments. SKH1 mice were chosen for this study because they mimic human hair status at the dermis, which is critical in dermal exposure surrogate/subcutaneous translational models. SKH1 mice are an important translational model for cutaneous and subcutaneous applications, from drug discovery to toxicity studies, as they are euthymic and immunocompetent, whereas other nude mice are not³³. SKH1 mice are ideal for imaging applications^14–19 because they do not require depilation before assaying. Subcutaneous exposure in SKH1 mice serves as a surrogate model for human dermal exposure or missed subcutaneous injection, which are growing public health and occupational safety concerns for fentanyl^20–22. Unfortunately, replacement/alternative models of systemic cardiovascular and cardiopulmonary toxicity are not currently available, but animal numbers were kept minimal to satisfy power analysis based on previous data^34,35. Experiments and analyses were performed unblinded due to accepted procedures of the up-down procedure, which seeks to identify the critical features of the dose-response curve while utilizing minimal animals³⁵.

Modified Up-Down Procedure

A modified up-down procedure was performed for fentanyl citrate dissolved in standard Ringer’s solution. In a typical up-down procedure³⁵, a literature LD50 dose is administered to two animals. If the dose produces 0% mortality, it is raised by one semi-log (3.2x), or lowered 3.2x after 100% mortality. Further doses are titrated between these values to generate at least four different doses to generate an LD50 curve via 4-parameter logistic regression. The initial dose in this study was based upon the limited LD50 studies in mice was 62 mg/kg²³, but failed to produce mortality. One semi-log higher dose would have required a larger injection volume than generally advised³⁶ due to the solubility of fentanyl (25 mg/mL)³⁷; therefore, a maximum dose was set at 160 mg/kg. The up-down procedure was then followed normally for doses between 62 mg/kg and 160 mg/kg. Due to the utilization of this procedure to limit animal numbers, randomization of doses was not permitted.

Drug Preparation

Fentanyl citrate (Spectrum Chemical, Gardena, California; #F1147, lot 1JC0116, 99.2% purity) was prepared as stock solution at 25 mg/mL in mammalian Ringer’s solution (Fisher Scientific, Waltham, Massachusetts; S25513, lot 0GI20080611A) as a vehicle then further diluted to volumes of 80–200μL. Empty vehicle (200μL) was used in control animals to confirm no effect from dose volume. All procedures involving fentanyl citrate were performed using fentanyl-resistant gloves (Ansell MICROFLEX^™ MIDKNIGHT^™ XTRA 93–862 Nitrile Gloves with Extended Cuff; Ansell, Richmond, Australia). All lab members were trained in Narcan (Adapt Pharma, Radnor, Pennsylvania) administration and a second lab member was present for all fentanyl handling in case of accidental exposure.

Drug Administration and Animal Monitoring

Animals were anaesthetized under 2% isoflurane (VetOne, Boise, Idaho) in a vented chamber then moved to a nosecone with 1.5–2.5% isoflurane. After confirming normal vitals (pulse, SpO2, temperature, and respiratory rate), shallow subcutaneous injections were performed using 31g syringes in the area behind the neck, between the shoulders and ears of the animals. Isoflurane was immediately removed after fentanyl administration. Pulse oximetry data were collected at 10-second intervals for five minutes post-injection with a tail clip pulse oximeter (Patterson Veterinary, Devens, Massachusetts). All mice were monitored immediately post-exposure for the first 30 minutes, then at least once every 30 minutes for the first six hours for signs of distress and recovery. Animals showing signs of mortality (e.g. absence of detectable output via pulse oximetry or plethysmography) were immediately humanely euthanized under 4–5% isoflurane. Surviving animals were monitored twice daily for seven days post-injection.

Whole Body Plethysmography

A Buxco FinePointe Whole Body Plethysmography 2-Site System was used with FinePointe software (version 2.4.6.9414) for whole body plethysmography (Data Sciences International, St. Paul, Minnesota). Baseline plethysmography data were recorded for at least 45 minutes before fentanyl treatment and post-exposure data were collected starting five minutes post-injection and continued for 30 minutes. The data collection epoch was left at its default value of two seconds. In order to reduce false measurements (e.g., animals moving around in recording chambers), recordings were filtered to only include breaths per minute (bpm) less than 385³⁸ with less than 10% change from the average of the preceding two epochs. Data were then binned into 5-minute segments and post-injection values were corrected to a baseline value consisting of the final 5-minute bin before injection.

Positron Emission Tomography

Animals were placed on a heating pad for 30 minutes prior to injection of [18F]Fluorodeoxyglucose (¹⁸F-FDG) to reduce fat uptake. An injection of 50–150 μCi of ¹⁸F-FDG was then administered via tail vein. Mice were immediately placed under isoflurane on a heating pad inside an enclosure for 20 minutes in order to reduce ¹⁸F-FDG muscle uptake. After this uptake period, the animals were placed transferred to an enclosure specially-designed for imaging in the PET scanner³⁹. In addition to having input and output ports for circulation of isoflurane, the enclosure has an integrated heated waterbed. The enclosure was placed in the PET scanner. They were scanned for 5 minutes. The resulting list mode data were reconstructed using the ordered set expectation maximization (OSEM) algorithm. Following completion of the scan, the animals were euthanized via cardiac puncture. After necropsy, radioactivity concentrations in the organs were measured using a well counter (Capintec 55tw). Organ data were time-corrected for radioactive decay and normalized to liver radioactivity. PET scans were performed on animals given no fentanyl (controls) or at various time points post-fentanyl injection (40 minutes, 6 hours, 24 hours, 7 days) to determine the time course of glucose uptake (a surrogate of inflammation) in response to fentanyl.

Statistical Analysis

Mortality data were analyzed using a four-parameter logistic regression with 0% survival as the lower bound and 100% survival as the upper bound in Prism 9.0 (GraphPad, San Diego, California). Pulseox data were binned into 30-second segments then analyzed via two-way ANOVA compared to pre-injection data within each respective treatment group. All pulseox data met the assumptions associated with two-way ANOVA as tested by Prism 9.0. Plethysmography data were analyzed using a mixed-effects model using the Geisser-Greenhouse correction and to compare between groups. All plethysmography data met the assumptions of mixed-effects model analysis and the Geisser-Greenhouse correction was used as a conservative measure. No data inclusion/exclusion criteria were used during analysis. All subjects were included in all analyses except for two plethysmography control datasets due to equipment error. Significance was set to p<0.05 for all analyses. All values are expressed as mean +/− standard error of the mean (SEM).

Results

Initial subcutaneous fentanyl exposures at the literature LD50 dose (62 mg/kg) produced zero mortality in SKH1 mice. In accordance with up-down procedures³⁵, we sought to raise the dose by semi-log of that LD50, to 198.4 mg/kg. However, due to the aqueous solubility of fentanyl citrate (25 mg/mL)³⁷ and the limitation of subcutaneous injection volume in mice³⁶ (5 mL/kg; 200μL for a 40g mouse), we could only raise the dose to 160 mg/kg. This dose initially produced 100% mortality in two animals. We then continued to use intermediary doses between 62 and 160 mg/kg to establish our LD50 curve (Figure 1). Additional animals were treated with fentanyl to verify lethality; notably, the 160 mg/kg dose produced 89% mortality (8/9 animals) instead of 100% mortality as initially observed in two animals. All animals given vehicle (n = 5) or 62 mg/kg fentanyl (n = 12) survived. Fentanyl doses that caused lethality were 110 mg/kg (1/10), 135 mg/kg (5/10), and 160 mg/kg (8/9). A small group of female animals were exposed to these doses of fentanyl (2 per dose) to explore sex-dependence of lethality rates: 0/2 at 62 mg/kg, 0/2 at 110 mg/kg, 1/2 at 135 mg/kg, and 2/2 at 160 mg/kg.

Figure 1.

Subcutaneous fentanyl mortality. Fentanyl citrate in Ringer’s solution was administered to SKH1 mice in the subcutaneous space behind the neck using an up-down procedure. All animals given vehicle (n = 5) or 62 mg/kg (n = 12) survived. Mortality was observed at doses of 110 mg/kg (n = 10), 135 mg/kg (n = 10), and 160 mg/kg (n = 9). Parentheses show n (n mortality / n total number of mice) for each dose. Four parameter logistic curve fitting closely matched (r² = 0.9996) observed mortality rates, with an LD10 dose at 111.2 mg/kg and an LD50 dose at 134.6 mg/kg. The Hill Slope of the curve was calculated to be 11.49.

For one minute before and five minutes after administration of either fentanyl or vehicle, pulse oximetry (HR and SpO2) was recorded at an epoch of 10 seconds and averaged over 30-second intervals for five minutes (Figure 2). No significant differences in HR or SpO2 were found between the doses that were capable of causing lethality (110–160 mg/kg), and therefore the responses from these mice were grouped together as overdoses. All overdose fentanyl animals exhibited a drop in SpO2 within 5 minutes. Overdose subjects naturally fell into three groups based on mortality and SpO2: survival (n = 15), non-survival with minimum SpO2 ≥37% (n = 8), and non-survival with minimum SpO2 <37% (n = 6). All mice that experienced a drop in SpO2 below 37% (n = 6) died within 35 minutes of injection, with 50% dying within seven minutes. For non-surviving overdose mice whose SpO2 minimum was at or above 37%, the average time of death post-dose was 44.9±7.4 minutes.

Figure 2.

Pulse oximetry before and after fentanyl overdose. (A) displays changes in SpO2 and HR in surviving animals (n = 15), (B) displays changes in SpO2 and HR in non-surviving animals whose SpO2 minimum was ≥37% after injection (n = 8), and (C) displays changes in SpO2 and HR in non-surviving animals whose SpO2 fell under 37% (n = 6). Vehicle-treated controls are displayed in blue in all graphs (n = 3). Data were averaged into 30-second bins for analysis. Vertical dotted line at t = 0s indicates time of injection; vertical marks above x-axis indicate the first significant (p<0.05) change in SpO2 or HR compared to its respective pre-injection value for fentanyl-treated animals. Arrows beginning at SpO2 and HR markers on x-axis indicate significance (p<0.05) at all time points beyond the initial drop compared to pre-injection. Two-way ANOVA was used to compare results within control or treated animal groups to pre-injection values. Summary statistics are provided in Supplemental Tables 1 and 2.

Control animals (administered 200μL Ringer’s solution subcutaneously) showed no significant difference in SpO2 or HR compared to pre-exposure values (Figure 2). From the overdose groups, surviving mice administered fentanyl showed a significant decrease in HR at 50 seconds followed by a significant decrease in SpO2 at 80 seconds, yielding an average 30-second lag between heart changes and distally measured (tail vein) systemic oxygen impacts (Figure 2A). Non-surviving fentanyl overdose mice whose SpO2 minimum was ≥37% first showed a significant decrease in HR at 110 seconds, followed by a significant decrease in SpO2 at 140 seconds, yielding a 30-second delay between heart and systemic oxygen effects of fentanyl (Figure 2B). HR did not significantly decrease in overdosed mice whose SpO2 went below 37%, but SpO2 was first significantly lower than baseline at 110 seconds (Figure 2C) post-exposure to fentanyl. The lack of significant HR response is unique to the <37% SpO2 group and this cardiovascular dysregulation may reflect muscle rigidity linked to WCS, yet this mechanism requires further elucidation.

Mice were introduced to plethysmography chambers for an acclimation period of at least 45 minutes before any exposure in order to record baseline respiration data. They were subsequently removed from the chambers, administered subcutaneous fentanyl or vehicle, monitored for five minutes via pulse oximetry as stated above, then placed back in the chambers for at least 30 minutes of recording. Tidal volume (amount of gas respired per breath (Figure 3A), breaths per minute (Figure 3B), minute volume (amount of gas respired per minute; Figure 3C), and inspiratory quotient (the ratio of inspiratory time to expiratory time; Figure 3D) were calculated in 5-minute bins for animals given vehicle, animals given lethal doses of fentanyl (110–160 mg/kg) that survived, and non-surviving mice given lethal doses of fentanyl. Mice whose SpO2 decreased below 37% were not included in these analyses due to non-survival.

Figure 3.

Plethysmography results for SKH1 mice exposed to fentanyl. Data were acquired from mice administered vehicle (n = 5), administered fentanyl (110–160 mg/kg) that survived (n = 15), and non-surviving mice administered fentanyl (n = 8). (A) displays tidal volume, (B) displays breaths per minute, (C) displays minute volume, and (D) displays inspiratory quotient (the ratio of inspiratory time to expiratory time). Data were corrected to baseline values consisting of the average of 5 minutes of recording immediately before injection. * indicates significance (p<0.05) between Survival and Non-survival groups; X indicates significance (p<0.05) between Vehicle and Survival groups; + indicates significance between non-survivor and vehicle groups. Significance values between groups was calculated in a mixed-effects model using the Geisser-Greenhouse correction. Raw plots of minute volume data are provided in Supplemental Figure 1. Summary statistics are provided in Supplemental Tables 3–6.

Tidal volume (Figure 3A) was initially decreased (5 minutes; p<0.05) between controls and each of the surviving and non-surviving fentanyl groups. While the non-survival group recovered thereafter (not significantly different compared to controls), the survival group was decreased (p<0.05) compared to controls at minutes 10, 25, and 30. Breathing rate (bpm; Figure 3B) was decreased for both survival and non-survival mice at 5 minutes post-fentanyl when compared to controls; it remained decreased (p<0.05) for non-survival mice compared to survivors across all time points measured. Minute volume (Figure 3C) was significantly decreased between non-survivors and controls for minutes 5, 10, and 35 post-fentanyl administration, but non-survivors had decreased (p<0.05) values compared to survivors at all time points. Similar trends appear in inspiratory quotient (Figure 3D), where non-survivors had decreased (p<0.05) values for the initial 5–15 minutes compared to controls, while non-survivors were significantly decreased when compared to survivors across nearly all time points. By 40 minutes post-exposure to lethal doses of fentanyl (110–160 mg/kg), there were no significant differences (p<0.05) in any breathing parameters between control and surviving animals.

In order to investigate general whole-body metabolic effects of lethal, high-dose fentanyl, ¹⁸F-FDG PET scans were performed on mice beginning at 40 minutes post-injection. We selected the observed LD10 value in order to represent an overdose while increasing the opportunity for animal survival through the 7-day period. Representative images from animals (Figure 4) were reconstructed to exclude the bladder and kidneys, as high concentrations of pooled ¹⁸F-FDG there attenuates relative signal intensity in other organs. Brains and hearts are readily visible in control animals (example in top row), yet lungs were difficult to identify due to their size, signal intensity from hearts, and low overall ¹⁸F-FDG uptake. Methodology was employed to reduce muscle and fat uptake of ¹⁸F-FDG (see Materials and Methods), yet some uptake is visible in the shoulders and forelimbs in some animals. Each row of images represents a single animal.

Figure 4.

Representative effects of fentanyl via ¹⁸F-FDG PET imaging. SKH1 mice received vehicle or LD10 fentanyl at the time indicated before imaging. Each row indicates a single representative animal. Left column shows sagittal view aligned at the center of the heart; middle and right columns display prone and supine views focused on the centers of the heart and brain, respectively.

Immediately following PET imaging, mice were sacrificed and organs were isolated for residual ¹⁸F-FDG activity measurement in a well counter. Per standard radiology procedures, organ activity was time-adjusted for decay from the time of injection, normalized to organ weight, then normalized to liver activity^40–42 (Figure 5). Two animals were used per time point with the exception of 40 minutes post-fentanyl, which represents three mice. Average ¹⁸F-FDG uptake was reduced in all organs at time points immediately post-injection (40 minutes, 6 hours). Pre-injection ¹⁸F-FDG uptake appears to have returned to baseline by 24 hours, though with some potential overcompensation in the heart. By one week post-fentanyl exposure, ¹⁸F-FDG uptake remained stable.

Figure 5.

Residual activity in organs post-PET scan. Organs were harvested immediately post-imaging then residual radioactivity was measured in a well counter, with counts adjusted account for variability in radiotracer dosage^40–42. Time points on the x-axis represent time post-fentanyl injection; control indicates animals administered vehicle. Each symbol represents a single animal; two animals were used per time point with the except of 40 minutes post-fentanyl, which represents three animals. Summary statistics are provided in Supplemental Table 7.

Discussion

In this study, SKH1 mice were exposed to lethal fentanyl doses in order to describe physiological effects of overdoses. This work is the first to assess the subcutaneous LD50 of fentanyl in the SKH1 mouse model, overdose survival with regard to cardiovascular responses in the first five minutes, respiratory responses within the first forty minutes, and glucose uptake up to one week after single bolus administration of fentanyl. These findings demonstrate both the immediate and longer-term impacts physiological impacts, and physiological parameters correlated to survival.

Initial experiments revealed greater resistance to lethal fentanyl overdose in SKH1 mice than previous subcutaneous overdose studies. The primary literature subcutaneous LD50 dose of fentanyl, 62 mg/kg, was measured in Swiss-Webster mice in 1964²³ and is frequently referenced in Safety Data Sheet (SDS) documentation, however two subsequent studies in Swiss-Webster mice reported subcutaneous LD50 values of 7.57²⁴ and 28.5 mg/kg²⁵. These LD50 doses are less than half the measured LD50 in the present study, at 135 mg/kg. SKH1 mice are generally distinguished by a retroviral integration in their Hr gene, causing hairlessness with immunocompetence⁴³. Other differences between Swiss-Webster and SKH1 mice have not been reported beyond skin permittivity⁴⁴. While subcutaneous fentanyl lethality has not been explored in other mouse strains, it has been sparsely reported via other routes of administration. Intravenous LD50s to range from 6.9 mg/kg in Swiss Albino⁴⁵ mice to 11.2 mg/kg in Kunming mice⁴⁶. Intraperitoneal injection LD50 has been reported to be less than 3 mg/kg in Swiss Webster mice⁴⁷, 17.5 mg/kg in Swiss Albino mice⁴⁸, and 131.3 mg/kg in C57BL/6 mice⁴⁹. The oral LD50 is reported to be 368 mg/kg in MSDS literature without reference to strain⁵⁰ and was found to be 27.8 mg/kg in Swiss Albino mice⁴⁵. Subcutaneous exposure may produce similar variability among strains; for comparison, one study using subcutaneous morphine reported an LD50 range of 212–882 mg/kg across eight strains of mice⁵¹, but did not include SKH1 mice.

Previous non-invasive research into the effects of fentanyl on respiration has included blood oxygen monitoring, typically via pulse oximetry or direct measurement from blood samples, and plethysmography, which measures breathing as a function of gas volume changes in a closed space. Direct and indirect blood oxygenation measurements have been shown to decrease after fentanyl administration^52–54, as have HR, which has a direct relationship with blood oxygenation under normal circumstances^55,56. However, these studies have only investigated low doses, such as those seen with anaesthetic regimens, which are not lethal and are not representative of physiological parameters associated with overdose as explored herein. In this study, lethal doses of fentanyl (110–160 mg/kg) representing experimentally validated LD10, LD50, and LD90 doses in SKH1 mice were not found to exhibit significant dose-dependent differences in pulse oximetry or plethysmography results, and were therefore combined into a single overdose group.

In this study, SKH1 pulse oximetry data from exposures to an overdose yielded both expected results based upon μ opioid receptor agonism^57,58 and a novel relationship between HR and SpO2 associated with fentanyl overdose. In surviving animals, HR decreased first, followed by SpO2 30 seconds later in line with OIRD⁵⁹. In comparison, non-surviving animals (whose overall SpO2 minimum was at or above 37%) had a delayed decrease in HR (when compared to survivors) followed immediately by a decrease in SpO2. These data align with previous findings that SpO2 decreases in response to fentanyl, yet our results show that an early cardiac response may be critical for survival. An unexpected finding was a group of non-surviving, <37% SpO2, animals that experienced a significant decrease in SpO2 without concomitant decrease in HR, demonstrating a potential non-opioid receptor-mediated lethality associated with cardiovascular dysfunction caused by fentanyl. These observations, coupled to literature associated with known symptoms, point to the presence of WCS in this mouse model^60,61, but this requires further verification. In humans, WCS is known to onset within minutes⁶¹, in line with our observations of mortality. Clinically, WCS is treated with a neuromuscular blocker such as succinylcholine, simultaneously with naloxone and endotracheal intubation; naloxone-only reversal of WCS requires very high doses and has limited efficacy⁶⁰.

Plethysmography revealed persistent decreases (p<0.05) between non-surviving and surviving animals exposed to fentanyl overdose for the entirety of recording in breaths per minute, minute volume, and inspiratory quotient, up to 40 minutes post-administration. With an average time of death 44.9±7.4 minutes post-exposure to fentanyl, these trends indicate vital determinants for survival. Our plethysmography data generally agree with previous reports in rodents, showing decreases in tidal volume, breaths per minute, minute volume, and inspiratory quotient within 10 minutes of fentanyl administrations^34,54,62,63. However, while previous rodent studies have shown dose-response relationships via plethysmography at low doses^34,64, we did not observe dose-response effects across the overdose range used in this study. This unique plateau effect of fentanyl overdose on breathing implies saturation of underlying μ receptor-mediated central respiratory depression and secondary effects determining survival. Data from other studies^65,66 suggest a plateau effect of fentanyl on minute volume, driven by depressed respiratory rate, in a cumulative dose model at doses below 3.2 mg/kg in Swiss Webster mice. While those studies do not report significant changes in tidal volume like the present study, differences in strain and dose may explain discrepancies. Receptor saturation may be an important determinant to qualify overdose clinically, coupled to prognostic SpO2, HR, and plethysmography thresholds, yet these require additional studies.

Fentanyl is described primarily as a μ-receptor agonist but also has significant affinity for non-opioid receptors^67,68 including adrenoreceptors (α_1B > α_1A > α_1D), dopamine receptors D₁ and D_4.4, serotonin receptor 5-HT_1A⁶⁷, and muscarinic receptor M₃⁶⁸. While these receptors are expressed in the brain, they are also enriched in many tissues and organs, including the vasculature, smooth and skeletal muscle, heart, and lungs⁶⁹. In the vasculature, α₁ receptors (particularly α_1D⁷⁰) induce vasoconstriction upon norepinephrine binding, while local M₃ receptor activation can induce vasodilation⁷¹. α_1A, α_1B, and M₃ receptors have prominent roles in the myocardium regarding contractility and ventricular function^70,71, and airway smooth muscle expresses α_1A, α_1B⁷², and M₃⁷³ as well. Further, fentanyl is known to be sequestered into the lungs directly after administration⁷⁴, potentially leading to sustained local effects. While brain region-specific effects of fentanyl on norepinephrine and acetylcholine release have been investigated¹², future studies should address circulating levels of these hormones in response to fentanyl due to the presence of these receptors throughout the vasculature and organ systems.

Few studies have investigated the effects of fentanyl in organs outside the brain. Physiological outcomes in the lungs^8,9 and heart^55,56 have been described and μ-receptors are known to be present in these and other organs^75–77. While hypoxia in these organs has been discussed at length⁷⁸, sustained hypoxia induced by fentanyl has not. Glycolytic oxygen consumption is a master regulator of global metabolism; therefore, we performed ¹⁸F-FDG PET scans in order to investigate the effects of fentanyl on organ glucose uptake at time points post-acute fentanyl administration (40 minutes, 6 hours, 24 hours, and 7 days). ¹⁸F-FDG is a β-emitting glucose isostere that is imported into cells via glucose transporters. PET imaging provided an organ map of relative signal intensity, while residual β emission activity in necropsied organs provided direct evidence of altered glucose uptake.

In typical hypoxia, glucose uptake has been shown to be broadly increased⁷⁹. However, our PET results show decreased ¹⁸F-FDG at 40 minutes and 6 hours post-administration, with a return to near baseline values at 24 hours in all organs assayed. These organs were normalized to residual liver activity measurements in order to standardize for variations in ¹⁸F-FDG dose^40–42; although the liver itself may be impacted by fentanyl, it is metabolized via CYP3A4 and is not known to directly cause liver injury⁸⁰. Glucose uptake is primarily controlled by GLUT1 (brain, blood-brain barrier), GLUT2 (kidneys, liver, pancreas), GLUT3 (brain, lungs), and GLUT4 (adipose, heart, skeletal muscle)^79,81,82. GLUT1 and GLUT4 are insulin-sensitive, and while a single study has shown fentanyl to decrease glucose-sensitive insulin release from rat pancreatic β-islets in vitro⁸³, this action alone may not explain reduced glucose uptake as seen in our study. Mice were fasted for 4 hours leading up to imaging and thus expected to be at basal circulating glucose and insulin levels, though concentrations were not confirmed. Decreased glucose uptake in our study is also confounded by lung results, as lungs have been shown to possess insulin-insensitive GLUT2 and GLUT3⁸¹. We did not expect decreased glucose uptake with fentanyl-induced hypoxia, yet these results show that the overall metabolic impact of overdose fentanyl is not limited to the initial 40-minute period of intoxication. While the effect of hypoxia on insulin levels has been described as highly variable⁸⁴, the impact of fentanyl on glucose uptake warrants further investigation.

This study defines a new subcutaneous LD50 in mice, provides physiological parameters that precede mortality with lethal doses of fentanyl within 40 minutes of administration, and displays impairment of glucose uptake in brains, hearts, and lungs up to 24 hours after administration. Our findings include (1) SpO2<37% after fentanyl injection is predictive of mortality, (2) an earlier, significant decrease in HR before a significant decrease in SpO2 in indicative of survival, (3) a lack of significant HR response coupled with a decrease in SpO2 post-fentanyl is associated with mortality, and (4) sustained, significant changes in breaths per minute, minute volume, and inspiratory quotient for minutes 5–40 after fentanyl administration stratify survival versus mortality in SKH1 mice. While SpO2 and HR data may reflect the onset of WCS in this model, given the association of HR with survival, further validation of WCS is necessary. The SpO2, HR, and plethysmography results from this study may serve as prognostic factors for use as study endpoints for animal protocol development and guide future Animal Care and Use policies for overdose research. The results from our PET studies show that glucose uptake is impaired in critical organ systems at time points beyond the initial period of fentanyl intoxication with recovery requiring 24 hours. While respiratory depression and other life-threatening effects of acute fentanyl overdose require immediate treatment, prolonged organ effects may require further treatment. Future research is necessary into both the short- and long-term effects of fentanyl in order to find novel methods to address its lethality and threat to quality of life.

Supplementary Material

Supinfo1

Supplemental Figure 1. Raw minute volume data used to generate Figure 3C.