Low-dose morphine reduces tolerance to central hypovolemia in healthy adults without affecting muscle sympathetic outflow

Abstract

Hemorrhage is a leading cause of preventable battlefield and civilian trauma deaths. Low-dose (i.e., an analgesic dose) morphine is recommended for use in the prehospital (i.e., field) setting. Morphine administration reduces hemorrhagic tolerance in rodents. However, it is unknown whether morphine impairs autonomic cardiovascular regulation and consequently reduces hemorrhagic tolerance in humans. Thus, the purpose of this study was to test the hypothesis that low-dose morphine reduces hemorrhagic tolerance in conscious humans. Thirty adults (15 women/15 men; 29 ± 6 yr; 26 ± 4 kg·m⁻², means ± SD) completed this randomized, crossover, double-blinded, placebo-controlled trial. One minute after intravenous administration of morphine (5 mg) or placebo (saline), we used a presyncopal limited progressive lower-body negative pressure (LBNP) protocol to determine hemorrhagic tolerance. Hemorrhagic tolerance was quantified as a cumulative stress index (mmHg·min), which was compared between trials using a Wilcoxon matched-pairs signed-rank test. We also compared muscle sympathetic nerve activity (MSNA; microneurography) and beat-to-beat blood pressure (photoplethysmography) during the LBNP test using mixed-effects analyses [time (LBNP stage) × trial]. Median LBNP tolerance was lower during morphine trials (placebo: 692 [473–997] vs. morphine: 385 [251–728] mmHg·min, P < 0.001, CI: −394 to −128). Systolic blood pressure was 8 mmHg lower during moderate central hypovolemia during morphine trials (post hoc P = 0.02; time: P < 0.001, trial: P = 0.13, interaction: P = 0.006). MSNA burst frequency responses were not different between trials (time: P < 0.001, trial: P = 0.80, interaction: P = 0.51). These data demonstrate that low-dose morphine reduces hemorrhagic tolerance in conscious humans. Thus, morphine is not an ideal analgesic for a hemorrhaging individual in the prehospital setting.

NEW & NOTEWORTHY In this randomized, crossover, placebo-controlled trial, we found that tolerance to simulated hemorrhage was lower after low-dose morphine administration. Such reductions in hemorrhagic tolerance were observed without differences in MSNA burst frequency responses between morphine and placebo trials. These data, the first to be obtained in conscious humans, demonstrate that low-dose morphine reduces hemorrhagic tolerance. Thus, morphine is not an ideal analgesic for a hemorrhaging individual in the prehospital setting.

INTRODUCTION

Hemorrhage is a leading cause of preventable battlefield and civilian trauma deaths (1–4). Surviving a hemorrhagic insult is highly dependent on immediate treatment in the prehospital (i.e., field) setting (5–7). In the early stages of hemorrhage-related hypovolemic shock, three primary compensatory mechanisms are vital to maintaining perfusion pressure, and thus, blood flow to vital organs (8, 9). First, there is parasympathetic inhibition to raise the heart rate. Second, there is upregulation of the renin-angiotensin-aldosterone system that increases plasma arginine vasopressin concentrations. Third, efferent sympathetic outflow increases to reduce vascular conductance, raise heart rate, and increase cardiac contractility. Eventually, individuals can transition into a state of noncompensable hemorrhage, with accompanied parasympathetic stimulation and sympathetic withdrawal. This leads to increased vascular conductance and reduced cardiac output until arterial pressure falls below a point necessary to perfuse vital organs (e.g., brain and heart) to meet metabolic demands, culminating in death (9). This progression occurs within minutes to hours following a hemorrhagic insult, which has important implications for the field care of civilians and soldiers in the prehospital setting.

Morphine (a μ-opioid receptor agonist) is the most frequently used analgesic for prehospital settings over the past century and is widely used in hospital settings (6, 10–17). Although recent US Army guidelines recommend oral transmucosal fentanyl citrate (OTFC) for pain management if an injured soldier is in moderate to severe pain and the individual is not in circulatory shock or respiratory distress, intravenous morphine is also indicated under such conditions if intravenous access has been obtained (18). Despite this recommendation, there is a paucity of data regarding whether low-dose morphine (5 mg, as used in the field setting) affects hemorrhagic tolerance in conscious humans. Importantly, an understanding of how low-dose morphine affects vital physiological processes, such as autonomic nervous system control of blood pressure (BP) (19, 20), would improve clinical risk-benefit analyses for its use in conscious humans.

A prior study found that morphine administration severely reduces hemorrhagic tolerance, in a dose-dependent manner, in conscious rodents (21). However, that observation is in contrast to a more recent investigation where morphine did not reduce hemorrhagic tolerance in conscious rodents (22). Such detrimental effects, when present (21), may be related to morphine depressing ganglionic transmission, as demonstrated in rabbits (23), or reducing sympathetic nervous system-induced tachycardia as demonstrated in dogs (24). Also, morphine accelerates hemorrhage-associated hypotension in conscious sheep [intracerebroventricular infusion (25)] and anesthetized rodents [microinjection into the cuneiform nucleus (26)]. In humans, higher doses (15–20 mg) of morphine can cause fainting during the head-up tilt (27), potentially because of reductions in vascular resistance (28). Furthermore, chronic μ-opioid receptor stimulation is associated with depressed sympathetic neural and cardiovascular regulation during sodium nitroprusside-induced hypotension (29). However, no studies to date have examined whether low-dose morphine administration affects tolerance to central hypovolemia nor the associated autonomic cardiovascular responses in conscious humans. Therefore, the purpose of this study was to test the hypothesis that intravenous morphine administration reduces tolerance to a simulated hemorrhagic insult in conscious humans. Moreover, we used electrocardiography, photoplethysmography, and microneurography to comprehensively assess autonomic cardiovascular responses during simulated hemorrhage after morphine or placebo administration. The results from this work are highly clinically relevant and will inform treatment strategies (e.g., US Army Combat Casualty Care guidelines) for hemorrhage with a commonly used “field” analgesic (6, 10–17). Furthermore, these data will add novel insight into the effects of morphine on the complex integrative autonomic cardiovascular physiological responses to profound central hypovolemia.

METHODS

Ethical Approval

The US Army Medical Research & Material Command Human Research Protection Office and the Institutional Review Boards for Human Subjects Research at the University of Texas Southwestern Medical Center (IRB No. 092017-070) and Texas Health Presbyterian Hospital Dallas approved this protocol and the associated informed consent. This study conformed with the standards set by the latest revision of the Declaration of Helsinki. Each of the 44 participants provided verbal and written consent before enrollment in the study. The data in this manuscript are associated with a registered clinical trial (ClinicalTrials.gov identifier: NCT04138615).

Experimental Overview

This experiment with low-dose morphine was a randomized, crossover, double-blind, placebo-controlled design consisting of two experimental visits separated by at least 48 h. The experimental visits included initial pain assessments (before drug administration), a drug/placebo administration, a progressive lower-body negative pressure (LBNP) test, and a second pain assessment (after drug administration; see timeline in Fig. 1). The data from this protocol aimed to address two distinct research questions, i.e., the effect of low-dose morphine administration on 1) tolerance to progressive LBNP along with the integrative sympathetic and cardiovascular responses, from which data are presented herein and 2) pain perception, sympathetic, and cardiovascular responses during the cold pressor test (CPT), which are presented in a companion manuscript (Watso et al., under peer review at the date of this publication).

Figure 1.

Experimental design. The two experimental visits for each participant were identical except for one visit they received morphine and the other visit they received placebo (order randomized, participant blinded to trial day). Following instrumentation and a quiet rest period, we collected preinfusion baseline data. We next completed pain assessments (algometry and the cold pressor test). After a recovery period, we infused 5 mg of morphine or saline (placebo) intravenously over 60 s. Sixty seconds after the infusion was complete, we started the lower-body negative pressure (LBNP) test—see methods for details. After another recovery period, individuals repeated the pain assessments before discharge from the laboratory.

Participants

We recruited participants from the Dallas-Fort Worth metroplex. All participants who enrolled in this study were free of any known cardiovascular, metabolic, or neurological diseases. Inclusion criteria included: age between 18 and 45 yr, body mass ≥60 kg, and body mass index <35 kg/m² at screening. Exclusion criteria included current or recent (within the past 3 yr) nicotine use, and current use of pain modifying or antihypertensive medications.

Experimental Protocol

Before the experimental visits, participants were instructed to not take in water intake for 2 h; food for at least 6 h; caffeine for 12 h; avoid strenuous exercise, alcohol, and NSAID’s for 24 h; and avoid over-the-counter cold or allergy medication and aspirin for 36 h before each trial. Participants arrived at the laboratory and provided a urine sample to confirm 1) a negative urine drug screen, 2) euhydration with a spot urine specific gravity ≤1.025 (30) (Atago Inc., Bellevue, WA), and 3) a negative pregnancy screen for females. After an intravenous catheter was placed, participants were instrumented for the measurement of muscle sympathetic nerve activity (MSNA), brachial BP, beat-to-beat BP, heart rate, middle cerebral artery blood flow velocity (MCAv), cerebral oxygen saturation, respiration rate, and partial pressure of end-tidal carbon dioxide. Once these signals were acquired, participants relaxed quietly in the supine position in our temperature-controlled laboratory.

Following a 5-min epoch for baseline data collection and blood sample collection, participants completed a pressure pain threshold test and a CPT, which addressed unrelated hypotheses with those data presented in a companion manuscript. Then, for 60 s, we intravenously administered either saline (0.9% NaCl) or 5 mg of morphine (morphine sulfate, Hospira, Inc., Lake Forest, IL). This dose of morphine was selected based on the 2014 Tactical Combat Casualty Care Guidelines (18). Sixty seconds after drug/placebo administration, participants completed a progressive lower-body negative pressure (LBNP) test.

LBNP Protocol

The LBNP protocol started at 40 mmHg of negative pressure, with 10 mmHg of additional negative pressure applied every 3 min (e.g., −40 mmHg for 3 min, −50 mmHg for 3 min, −60 mmHg for 3 min, etc.) until the participant reached presyncope. If a participant did not reach presyncope during the 3 min at −100 mmHg, this pressure was maintained for up to an additional 5 min [i.e., maximum of 26 min = (7 stages × 3 min) + additional 5 min]. We terminated LBNP if at least one of the following criteria were met: 1) the subject had continued reports of feeling faint and/or nauseated, 2) there was a rapid decline in BP resulting in systolic BP less than or equal to 80 mmHg, and/or 3) there was relative bradycardia accompanied with narrowing of pulse pressure. We quantified hemorrhagic tolerance using the cumulative stress index (mmHg·min). This index is determined mathematically by summing the product of the negative pressure and time, in minutes (or fraction of a minute), for each stage [e.g., (40 mmHg ×·3 min) + (50 mmHg × 3 min) + (60 mmHg ×·3 min), etc.] until test termination (31). LBNP causes a redistribution of blood from the upper body to the lower body and elicits central hypovolemia (8). For this reason, progressive LBNP is a validated model of human hemorrhage that can be simulated in the laboratory setting (32, 33). Notably, LBNP tolerance is highly reproducible (34). In addition, neither LBNP tolerance nor LBNP-induced cardiovascular responses differ between menstrual cycle phases in young female adults (35, 36). Thus, we did not control for the menstrual cycle in all female participants. Nonetheless, we were able to match the menstrual cycle phase between trials in most female adults despite scheduling/logistical challenges related to performing human research during the SARS-CoV-2 pandemic (i.e., this study was conducted between October 2019 and January 2022). Namely, 11 female adults completed both trials in the same menstrual cycle phase (n = 4 follicular, n = 4 luteal, n = 3 using intrauterine devices who were not menstruating the day of testing) compared with only three who completed trials in different menstrual cycle phases (information from one individual was not reported).

Muscle Sympathetic Nerve Activity

As previously described (37–39), we directly recorded MSNA using ultrasound-guided radial microneurography in a subset of adults (see results text for sample sizes for each time point) from which we could obtain adequate MSNA recordings during LBNP in both visits. Briefly, with real-time ultrasound imaging, we inserted a tungsten recording microelectrode into a radial nerve of the upper arm and inserted a reference microelectrode ≤3 cm from the recording electrode. The electrical signal was amplified (×80–90,000), bandpass filtered (700–2,000 Hz), rectified, and integrated (time constant, 0.1 s) using a nerve traffic analyzer (Nerve Traffic Analyzer, model 662c-4; University of Iowa, Bioengineering, Iowa City, IA).

Cardiovascular Measures

We used single-lead ECG to continuously assess heart rate (GE Medical Systems, WI). We used photoplethysmography to assess beat-to-beat BP, as well as Modelflow-derived cardiac output (Finometer; Finapres Medical Systems, The Netherlands) at rest and during LBNP (40). Briefly, Modelflow is a three-element model that uses the arterial characteristic impedance, arterial compliance, and peripheral resistance to compute valid estimates of stroke volume (41). Thus, Modelflow allows for valid estimates of cardiac output (42). On a beat-to-beat basis, we defined BP waveform peaks as systolic BP, nadirs as diastolic BP, and the average value of the integrated BP waveform as mean BP. Total vascular conductance was calculated by dividing cardiac output by finometer-derived mean BP. To assess brachial BP, we used a microphone placed over the brachial artery to detect Korotkoff sounds triggered from the ECG signal (Tango M2 Stress Test Monitor, SunTech Medical).

Cerebrovascular and Respiratory Variables

We measured MCAv on the right side of the body using transcranial Doppler (Spencer Technologies, Redmond, WA). We measured respiratory rate and the partial pressure of carbon dioxide from expired air using a nasal cannula connected to a capnograph (9004 Capnocheck Plus; Smiths Medical International Ltd, Watford, UK). We also measured cerebral tissue oxygen saturation using near-infrared spectroscopy (Moor Instruments Inc., Wilmington, DE) (43).

Compensatory Reserve Index

We used a finger pulse oximeter with capabilities to extrapolate continuous beat-to-beat recordings of arterial blood flow waveforms and determine the physiological reserve to compensate for reductions in central blood volume (Flashback Technologies, Boulder, CO). Subsequently, we quantified compensatory reserve using a previously described algorithm, resulting in a compensatory reserve index (CRI) (44). Briefly, this index is a normalized value on a scale of 0 to 1, with “1” reflecting the maximal capacity of the summation of compensatory physiological mechanism required during central blood volume deficits and “0” implicating imminent cardiovascular instability and decompensation (45).

Data Analysis

Data were collected at a sampling rate of 625 Hz using Biopac (MP150, Biopac, Santa Barbara, CA). While blind to condition, an experienced investigator (J.C.W.) (37–39, 46, 47) visually inspected the sympathetic neurogram on a beat-to-beat basis to determine the presence/absence of MSNA bursts using Ensemble (Elucimed, Wellington, New Zealand). MSNA analysis was conducted per recent guidelines (48) using the following criteria: 1) >3:1 signal-to-noise ratio, 2) burst morphology consistent with MSNA bursts, and 3) a pulse-synchronous signal. MSNA was quantified as burst frequency (bursts·min⁻¹) and burst incidence (bursts·100 heart beats⁻¹). MSNA burst amplitude and area were not assessed because we could not confirm that the position of the microelectrode was the same for both trials. In other words, MSNA burst amplitude and area varied within participants and between visits given the repeated-measures design. For example, we recorded MSNA from opposing arms on separate occasions because of the proximity of experimental trials, coupled with LBNP-related shifts in the neurogram. Therefore, it would be inappropriate to estimate the MSNA burst amplitude or area for this protocol given that, even with burst amplitude and area normalization, alterations in baseline characteristics of the neurogram have a large influence on the resultant LBNP-related changes in burst amplitude and area (48, 49).

We obtained venous blood samples to assess circulating catecholamines before drug/placebo administration, during the final minute of the 50 mmHg LBNP stage, and immediately after LBNP test termination. These samples were collected in Lithium Heparin spray-coated tubes (BD Vacutainer, Oakville, ON). We stored these tubes on ice until centrifugation (2,000 g for 10 min) less than 60 min later. Finally, we stored plasma samples at −80°C until shipping on dry ice to the laboratory (ARUP Laboratories, Salt Lake City, UT) for high-performance liquid chromatography assessment of plasma epinephrine and norepinephrine concentrations.

Statistical Analysis

We present values as means ± SD or medians [interquartile ranges] for data that are not normally distributed. We estimated sample size from a power analysis based on prior investigations of LBNP tolerance in our laboratory (50). Specifically, we set the power at 95% with an adjusted α of 0.015 and calculated an estimated sample size of n = 15. To enable comparisons for the primary variable of interest (LBNP tolerance quantified as cumulative stress index) within each sex, we doubled this estimated sample size. For all other comparisons (e.g., cardiovascular and sympathetic data), we pooled male and female adults together to interrogate the mechanisms (e.g., MSNA) underlying morphine-induced changes in LBNP tolerance. We used a Wilcoxon matched-pairs signed-rank test to compare cumulative stress index between morphine and placebo trials because the data failed the Shapiro–Wilk test for normality; this was repeated within each sex. We used a mixed-effects model [time (LBNP stage) × trial] to compare the cardiovascular, cerebrovascular, respiratory, and sympathetic variables at distinct time points between morphine and placebo trials to account for drop-outs due to individual variability LBNP tolerance. The time points selected were preinfusion (5 min from the quiet rest period before initial pain assessments and drug/placebo administration), postinfusion (the 60-s period after drug/placebo administration and before the start of LBNP), LBNP40 (the final 60-s period of 40 mmHg LBNP), LBNP50 (the final 60-s period of 50 mmHg LBNP), and LBNP-End (the final 20-s period before LBNP termination). When we observed a significant interaction, we performed post hoc multiple comparisons using Šídák's multiple comparisons tests. We used Fisher’s exact tests to determine whether signs/symptoms at presyncope differed between trials. We analyzed data using GraphPad Prism 9 (GraphPad Software Inc., La Jolla, CA). We had insufficient data to conduct a two-way ANOVA for plasma epinephrine concentrations because 16 blood samples had undetectable values (laboratory values reported as <10 pg/mL; n = 5 for preplacebo, n = 7 for premorphine, n = 3 for postplacebo LBNP50, n = 1 for postplacebo LBNP-End). Therefore, to gather insight, we assigned a value of 9.9 pg/mL for those undetectable values to enable the comparison between time points and trials using a Friedman nonparametric test (JMP, version 13.2.1, SAS Institute Inc., Cary, NC). Aside from the primary outcome variable (LBNP tolerance quantified as the cumulative stress index), we did not create a dichotomous line of significance/nonsignificance (51, 52). However, when P values were below 0.10, we considered that value along with the physiological relevance for each variable to conclude the result of a given comparison. Finally, we report the effect size, when appropriate, for the primary comparison of interest—LBNP tolerance quantified as the cumulative stress index between morphine and placebo trials—to aid with interpretation.

RESULTS

Thirty of the 44 adults who enrolled completed this trial. Nine individuals lost interest, decided not to continue, or stopped responding to investigators before the first experimental visit; two did not meet screening criteria; one discontinued because of sensations posttrial fatigue after their first experimental visit (placebo trial); one discontinued because they were uncomfortable with LBNP testing; and one male adult was excluded because we had already completed testing in 15 male adults by the time we could schedule their visits. The experimental protocol is detailed in Fig. 1. We provide participant characteristics in Table 1. Spot urine specific gravity upon arrival at the laboratory was not different between trials (placebo: 1.011 ± 0.006 vs. morphine: 1.011 ± 0.007; P = 0.93). Environmental conditions were not different between trials: temperature (placebo: 23.0 ± 1.3 vs. morphine: 23.3 ± 1.1°C, P = 0.22) and relative humidity (placebo: 40 ± 11 vs. morphine: 38 ± 13%, P = 0.51).

Table 1.

Participant screening information

n (female/male)	30 (15/15)
Age, yr	29 ± 6 [21–40]
Body mass, kg	77 ± 8 [62–97]
Body mass index, kg·m−2	26 ± 4 [20–34]
Systolic BP, mmHg	121 ± 9 [104–136]
Diastolic BP, mmHg	75 ± 8 [54–85]

Values are means ± SD [ranges]; n, number of subjects. BP, blood pressure.

One individual did not reach presyncope during their placebo trial within the time allotted, so the maximum cumulative stress index value possible (1,970 mmHg·min) was used to compare their respective morphine trial cumulative stress index value (1,207 mmHg·min). Nonetheless, morphine administration reduced LBNP tolerance (Fig. 2A). This morphine-induced reduction in LBNP tolerance was present in female and male adults when assessed independently as two separate groups (Fig. 2, B and C). There were no differences in signs (e.g., relative bradycardia) or symptoms (e.g., lightheadedness) between trials (P > 0.10 for all, data not shown).

Figure 2.

Tolerance to central hypovolemia. Morphine administration reduced lower-body negative pressure tolerance (A; Wilcoxon matched-pairs signed-rank test; median difference −209; confidence interval: −394 to −128). Morphine administration also reduced tolerance within female (B; Wilcoxon matched-pairs signed-rank test; median difference −159; confidence interval: −247 to −122) and male (C; paired, two-tailed t test; Cohen’s d = 0.70; mean difference −301; confidence interval: −538 to −64) adults when assessed as individual groups. We present data as median values with individual data for all participants and female adults. We present data as mean values with individual data for male adults.

The LBNP-induced reduction in systolic BP was 8 mmHg greater with morphine administration during LBNP40 (Fig. 3A). The LBNP-induced reduction in diastolic BP was 4 mmHg greater with morphine administration during LBNP-End (Fig. 3B). The LBNP-induced reduction in mean BP was 5 mmHg greater with morphine administration during LBNP40 (Fig. 3C). The LBNP-induced reduction in pulse pressure was 4 mmHg greater with morphine administration during LBNP40 (Fig. 3D). None of these variables were different between trials at any other time point.

Figure 3.

Blood pressure during progressive central hypovolemia. We compared systolic, diastolic, mean, and pulse pressure using mixed-effects analyses with both factors repeated. The lower-body negative pressure (LBNP)-induced reduction in systolic blood pressure (BP) was 8 mmHg greater with morphine administration during LBNP40 (time: F_4,116= 261, trial: F_1,29 = 2.4, interaction: F_4,90= 3.8; P > 0.11 for all other post hoc analyses) (A). The LBNP-induced reduction in diastolic BP was 4 mmHg greater with morphine administration during LBNP-End (time: F_4,116 = 160, trial: F_1,29= 1.9, interaction: F_4,90= 4.9; P > 0.27 for all other post hoc analyses) (B). The LBNP-induced reduction in mean BP was 5 mmHg greater with morphine administration during LBNP40 (time: F_4,116= 273, trial: F_1,29= 2.0, interaction: F_4,90= 4.4; P > 0.25 for all other post hoc analyses) (C). The LBNP-induced reduction in pulse pressure was 4 mmHg greater with morphine administration during LBNP40 (time: F_4,116= 162, trial: F_1,29= 0.72, interaction: F_4,90= 5.1; P > 0.24 for all other post hoc analyses) (D). The sample sizes were 28 for LBNP40 and 19 for LBNP50 due to reaching tolerance before the end of the respective stage in at least one of the two trials. We present data as means ± SD.

Heart rate values were 13 beats·min⁻¹ higher postinfusion and 7 beats·min⁻¹ higher during LBNP40 during the morphine trial (Fig. 4A). Heart rate values were 18 beats·min⁻¹ lower during LBNP-End during the morphine trial (Fig. 4A). Regardless of time point, the highest heart rate documented within each LBNP test was 11 beats·min⁻¹ lower during morphine trials (placebo: 123 ± 26 vs. morphine: 108 ± 31, P = 0.01). Stroke volume responses during LBNP did not differ between morphine and placebo trials (Fig. 4B). Cardiac output was 1 L·min⁻¹ greater postinfusion during the morphine trial (Fig. 4C). Total vascular conductance was 10 mL·min⁻¹·mmHg⁻¹ greater postinfusion during the morphine trial (Fig. 4D). The increases in MSNA burst frequency and burst incidence during LBNP were not different between trials (Fig. 5, A and B).

Figure 4.

Hemodynamics during progressive central hypovolemia. We compared heart rate, stroke volume, cardiac output, and total vascular conductance using mixed-effects analyses with both factors repeated. Heart rate values were 13 beats·min⁻¹ higher postinfusion and 7 beats·min⁻¹ higher during lower-body negative pressure (LBNP)40 during the morphine trial. Heart rate values were 18 beats·min⁻¹ lower during LBNP-End after morphine administration (time: F_4,116= 65, trial: F_1,29= 0.66, interaction: F_4,90= 25; P > 0.78 for all other post hoc analyses) (A). Stroke volume responses during LBNP did not differ between morphine and placebo trials (time: F_4,116= 142, trial: F_1,29= 0.092, interaction: F_4,90= 3.3; P > 0.22 for all post hoc analyses) (B). Cardiac output was 1 L·min⁻¹ greater postinfusion during the morphine trial (time: F_4,116= 52, trial: F_1,29= 2.2, interaction: F_4,90= 8.8; P > 0.95 for all other post hoc analyses) (C). Total vascular conductance was 10 L·min⁻¹·mmHg⁻¹ greater postinfusion during the morphine trial (time: F_4,116= 4.1, trial: F_1,29= 2.8, interaction: F_4,90= 2.8; P > 0.46 for all other post hoc analyses) (D). The sample sizes were 28 for LBNP40 and 19 for LBNP50 due to reaching tolerance before the end of the respective stage in at least one of the two trials. We present data as means ± SD.

Figure 5.

Sympathetic outflow during progressive central hypovolemia. We compared muscle sympathetic nerve activity (MSNA) burst frequency and incidence using mixed-effects analyses with both factors repeated. Increases in MSNA burst frequency were not different between trials (time: F_4,80= 37, trial F_1,20= 0.065, interaction: F_4,48= 0.84) (A). Increases in MSNA burst incidence were not different between trials (time: F_4,80= 12, trial: F_1,20= 0.098, interaction: F_4,48= 1.2) (B). The paired (between trials) sample sizes were 21 for pre/postinfusion, 19 for LBNP40, 12 for LBNP50, and 16 for LBNP-End due to signal dropout or individuals reaching tolerance before the end of the respective stage in at least one of the two trials. We present data as means ± SD.

The increases in plasma epinephrine concentrations during LBNP were not different between trials, though epinephrine concentrations were higher throughout the morphine trial (Table 2). The increases in plasma norepinephrine concentrations during LBNP were not different between trials (Table 2). Reductions in peak, mean, and minimum MCAv during LBNP were not different between trials (Table 2). Reductions in cerebral tissue oxygen saturation were not different between trials (Table 2). The respiratory rate was 3 breaths·min⁻¹ lower at LBNP-End within the morphine trials but did not vary between trials during other time points (Table 2). End-tidal CO₂ was 3 mmHg lower postinfusion lower within the morphine trials but did not vary between trials during other time points (Table 2). The compensatory reserve index was 0.2 units higher at LBNP-End within the morphine trials but did not vary between trials during other time points (Table 2).

Table 2.

Plasma catecholamine concentrations, cerebrovascular, and respiratory responses during central hypovolemia

	Preinfusion	Postinfusion	LBNP40	LBNP50		P Value
					LBNP-End	Time	Trial	Interaction
Epinephrine, pg·mL−1
n	21	n/a	n/a	10	21
Placebo	22 ± 15			65 ± 137	84 ± 78	<0.001	0.04	0.67
Morphine	36 ± 20			46 ± 35	123 ± 125
Norepinephrine, pg·mL−1
Placebo	220 ± 106			430 ± 174	988 ± 757	<0.001	0.85	0.67
Morphine	237 ± 79			488 ± 213	859 ± 923
Peak MCAv, cm·s−1
n	27	26	19	16	18
Placebo	90 ± 17	91 ± 16	81 ± 18	78 ± 16	65 ± 12	<0.001	0.82	0.11
Morphine	89 ± 20	87 ± 21	77 ± 18	75 ± 19	70 ± 17
Mean MCAv, cm−1·s−1
Placebo	62 ± 13	63 ± 11	56 ± 13	54 ± 12	39 ± 10	<0.001	>0.99	0.18
Morphine	62 ± 14	61 ± 15	54 ± 14	53 ± 14	43 ± 10
Minimum MCAv, cm·s−1
Placebo	43 ± 9	44 ± 8	41 ± 12	42 ± 11	24 ± 9	<0.001	0.59	0.13
Morphine	44 ± 10	43 ± 10	41 ± 11	41 ± 12	29 ± 8
Cerebral tissue oxygen saturation, %
n	26	26	24	16	26
Placebo	63 ± 7	64 ± 8	61 ± 9	59 ± 10	57 ± 10	<0.001	0.42	0.64
Morphine	65 ± 8	65 ± 9	62 ± 9	60 ± 11	59 ± 10
Respiratory rate, breaths·min−1
n	29	29	28	19	29
Placebo	13 ± 3	14 ± 4	13 ± 5	13 ± 5	16 ± 6	0.006	0.03	0.02
Morphine	14 ± 4	13 ± 4	12 ± 4	12 ± 4	13 ± 5a
End-tidal CO2, mmHg
n	30	30	28	19	27
Placebo	49 ± 5	48 ± 5	44 ± 6	42 ± 6	35 ± 7	<0.001	0.29	<0.001
Morphine	49 ± 5	45 ± 6b	42 ± 7	42 ± 7	37 ± 8
Compensatory reserve index, au
n	n/a	29	28	18	18
Placebo		0.9 ± 0.1	0.6 ± 0.2	0.5 ± 0.2	0.3 ± 0.2	<0.001	0.60	<0.001
Morphine		0.9 ± 0.1	0.6 ± 0.2	0.4 ± 0.2	0.5 ± 0.3c

Values are means ± SD. We used mixed-effects analysis to compare values. LBNP, lower-body negative pressure; MCAv, middle cerebral artery blood flow velocity; n/a, not applicable.

^aP = 0.004 between trials for LBNP-End; ^bP = 0.02 between trials during postinfusion; ^cP < 0.001 between trials for LBNP-End.

DISCUSSION

The primary novel finding of this investigation is that intravenous administration of low-dose morphine reduced tolerance to a simulated hemorrhagic insult in conscious humans. Importantly, the magnitude of the reduction in tolerance to central hypovolemia with morphine (>30%) is clinically meaningful. In addition, we found that systolic BP was 8 mmHg lower during moderate central hypovolemia during morphine trials. Such findings in conscious humans agree with most (21, 25), but not all (22), prior work in animals demonstrating detrimental effects of morphine on hemorrhagic tolerance and cardiovascular responses during hemorrhage. Related to compensatory mechanisms during central hypovolemia, it is possible that attenuated increases in “maximal” heart rate, as observed at LBNP-End, contribute to lower LBNP tolerance with morphine. The sympathetic nervous system is unlikely to have contributed to lower LBNP tolerance with morphine in the present work, as evidenced by a lack of differences in MSNA burst frequency between trials. Nonetheless, the present data support the CoTCCC guidelines that morphine should not be a primary option for hemorrhaging individuals in field settings (18).

Cardiovascular Responses to Hypotension and Central Hypovolemia

The present findings of reduced cardiovascular compensation (e.g., lower systolic and mean BP during moderate central hypovolemia) with a single low-dose morphine administration align with past findings of chronic μ-opioid receptor stimulation depressing sympathetic neural and cardiovascular regulation during sodium nitroprusside-induced hypotension (29). Related, prior work in patients with acute myocardial infarction demonstrated that morphine administration (3–10 mg) raised heart rate and reduced BP, with two of 12 patients having >30 mmHg drops in mean BP (11). Eight of these 12 patients experienced drops in peripheral resistance, which likely contributed to the cohort experiencing falls in BP. In agreement, we found that total vascular conductance was higher (i.e., reciprocally suggestive of lower resistance) postinfusion of morphine compared with placebo. Heart rate-mediated increases in cardiac output during this period may have been necessary to prevent BP from falling before the LBNP stressor started. Finally, in another study among healthy adults, intravenous morphine (15 mg) reduced forearm vascular resistance during 45° head-up tilt (another acute stimulus that elicits central hypovolemia) without consequent changes in mean BP (28). Discrepancies between this prior study with the present work, which did not find changes in vascular conductance during central hypovolemia but did find a 5 mmHg reduction in mean BP, could be due to differences in dosing (i.e., 5 vs. 15 mg). Taken together, our findings add to prior literature and demonstrate that a 5-mg administration of morphine has deleterious effects for cardiovascular responses during progressive central hypovolemia. Finally, a greater compensatory reserve index during LBNP-End after morphine administration, suggestive of less decompensation at presyncope, requires future investigation given the potential for this value to influence treatment decisions in field settings.

Sympathetic Nervous System Responses during Lower-Body Negative Pressure

Prior findings of morphine reducing hemorrhagic tolerance in animals (21, 25) may be related to morphine depressing ganglionic transmission, as demonstrated in rabbits (23), or due to reducing sympathetic nervous system-induced tachycardia, as demonstrated in dogs (24). Counter to the results from the former study (23), we did not find morphine administration to affect MSNA responses during LBNP (a sympathoexcitatory stimulus) in our cohort of healthy humans. In partial agreement with the latter study (24), we did identify that morphine reduced the peak heart rate during LBNP, but we were unable to determine whether the sympathetic nervous system mediated this attenuated tachycardic response. In a human study, the finding that naloxone (μ-opioid receptor antagonist) administration did not affect cumulative stress tolerated during LBNP, but did augment the peak increase in MSNA (53), suggested that morphine (μ-opioid receptor agonist) might attenuate increases in MSNA during LBNP. However, the current data demonstrate that the relation between μ-opioid receptor agonism and antagonism may not be of simple reciprocity. Finally, a previous report has described that morphine does not affect MSNA responses during other sympathoexcitatory stimuli including handgrip exercise and postexercise ischemia (54), which agrees with our finding of no morphine-induced changes in the MSNA response during LBNP. In summary, low-dose morphine administration did not affect MSNA burst frequency responses during progressive central hypovolemia.

Cerebrovascular Responses during Progressive Central Hypovolemia

Data from cats suggest that morphine, in concentrations similar to that which would occur in the present protocol in humans, decreased resting middle cerebral artery (MCA) diameter by ∼20% (55). Thus, given the strong influence of vessel diameter for the calculation of blood flow (Poiseuille’s Law), velocity findings from the MCA (e.g., MCAv) should be interpreted with caution given the potential for morphine to dilate the MCA, as demonstrated in cats. Importantly, our findings do not suggest that low-dose morphine affects MCAv at postinfusion and during progressive central hypovolemia. Regarding brain tissue oxygenation, a study in cats reported that morphine reduced global venous, but not arterial, oxygen saturation because of reduced oxygen consumption in the thalamus and hypothalamus (56). These past findings contrast with the present work, as we did not detect morphine administration to affect the magnitude of the reductions in cerebral tissue oxygen saturation during LBNP. The dissimilar results between studies are likely because of dosing and species differences. In conclusion, low-dose morphine administration did not affect MCAv nor cerebral tissue oxygen saturation in healthy adults challenged with progressive central hypovolemia.

Respiratory Responses during Progressive Central Hypovolemia

The US Army Combat Casualty Care guidelines recommend monitoring for respiratory depression following intravenous administration of morphine (18). Using the 5-mg dose recommended in these guidelines, respiratory rate was not different between trials, with the exception that morphine reduced respiratory rate by 3 breaths·min⁻¹ during the final 20 s of LBNP. Pertinent to this discussion, we identified only a 3 mmHg lower end-tidal CO₂ after morphine infusion and before LBNP started (i.e., see postinfusion in Table 2). The present findings are in agreement with a prior study using a higher dose of intravenous morphine (15 mg) that reported a 3–4 breaths·min⁻¹ reduction in respiratory rate (28). Although minute ventilation and tidal volume were not assessed, the present data suggest minor respiratory system depressant effects from a single low-dose morphine administration during LBNP.

Subjective Responses

In the present study, symptoms reported by participants during or immediately after LBNP did not differ between trials. This is in contrast to prior studies with reports of dizziness, feeling nauseated, blurry-eyed, lightheaded, sluggish, and fatigued to morphine itself (57–66). However, given that some of these sensations can occur with presyncopal-limited LBNP alone (e.g., dizzy, lightheaded), it is possible that the sensations of LBNP “out weighed” any sensations related to morphine administration. Also, the prior work tended to use higher doses of morphine than investigated here. Together, these reasons likely explain why participants in the present study did not report different symptoms during morphine and placebo trials.

Limitations

In the present work, we are unable to determine the extent to which morphine affected upregulation of the renin-angiotensin-aldosterone system (one compensatory mechanism that contributes to blood pressure/flow maintenance during central hypovolemia). Future studies are needed to evaluate the role of morphine in potentially affecting this hormonal axis during hemorrhage. Also, because low-dose morphine administration reduced BP during portions of LBNP testing (e.g., systolic BP lower at LBNP40, diastolic BP lower at LBNP-End), we are unable to determine whether morphine affects MSNA during central hypovolemia if that level of hypovolemia occurred without concurrent morphine-induced reductions in BP. Further studies are necessary to evaluate this question by clamping blood pressure in both the morphine and placebo trials, and then assessing MSNA responses during central hypovolemia. The pharmacodynamics/pharmacokinetics of morphine and morphine metabolites result in a ∼0.5-h delay until a sustained “maximal” analgesic effect is produced (59, 67–69), though such a maximal analgesic effect was not measured in the present study. Thus, it is unclear whether morphine would reduce hemorrhagic tolerance 30 or 60 min following drug administration. We chose to mirror the timing from prior studies using this experimental paradigm (38, 39) to allow for future comparisons between commonly used prehospital analgesics (which are well beyond the scope of this manuscript) and to maximize ecological validity. Namely, it is unlikely that morphine would be proactively administered minutes/hours before a hemorrhagic injury occurs. Nonetheless, future studies are warranted to examine whether morphine has sustained detrimental effects on autonomic cardiovascular regulation for hypovolemic and/or hypotensive states.

Perspectives and Significance

We found that low-dose morphine administration reduced tolerance to a simulated hemorrhagic insult in conscious humans. For context, the data from this cohort (mean body mass of 76.5 kg), coupled with findings from sedated baboons exposed to both blood loss and LBNP (33) suggest that appropriate cardiovascular compensation can occur until ∼1.5 L of blood loss after placebo administration (the median last LBNP stage completed was 70 mmHg). That finding contrasts with compensation only up to ∼1.1 L of blood loss after morphine administration and before presyncope (the median last LBNP stage completed was 50 mmHg). As mentioned, these calculations are based on prior data matching actual (blood removal) with simulated (LBNP stage) blood loss in sedated baboons matched for changes in central venous pressure and pulse pressure (33). Notably, there is also a strong relation in hemodynamic responses between actual and simulated blood loss in conscious humans (R² = 0.99 for central venous pressure and pulse pressure) (70). Though, we acknowledge that the blood removal data from baboons may not be completely applicable to conscious humans. Nevertheless, these findings suggest that morphine should be avoided for hemorrhaging individuals, particularly in prehospital (i.e., field) settings where advanced cardiopulmonary support equipment may not be immediately available. These data support current medical treatment guidelines for prehospital settings, primarily the notion that morphine is not the ideal choice for analgesia in a hemorrhaging individual (18, 71). Finally, these findings provide novel information about how low-dose morphine affects autonomic cardiovascular regulation during progressive central hypovolemia.

DATA AVAILABILITY

Data are available upon reasonable request to the principal investigator after institutional data transfer agreement approvals.

GRANTS

This research was supported by United States Army Department of Defense Grant W81XWH1820012 (to C.G.C.), American Physiology Society Postdoctoral Fellowship (to J.C.W.), and National Heart, Lung, and Blood Institute Grants F32HL154559 (to J.C.W.) and F32HL154565 (to L.N.B.).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

J.C.W., L.N.B., J.M.H., M.H., C.H.-L., and C.G.C. conceived and designed research; J.C.W., L.N.B., F.A.C., B.D.O., J.M.H., M.H., E.J., J.F., and C.G.C. performed experiments; J.C.W., F.A.C., and E.J. analyzed data; J.C.W., L.N.B., F.A.C., B.D.O., J.M.H., M.H., E.J., J.F., C.H.-L., and C.G.C. interpreted results of experiments; J.C.W. prepared figures; J.C.W. and C.G.C. drafted manuscript; J.C.W., L.N.B., F.A.C., B.D.O., J.M.H., M.H., E.J., J.F., C.H.-L., and C.G.C. edited and revised manuscript; J.C.W., L.N.B., F.A.C., B.D.O., J.M.H., M.H., E.J., J.F., C.H.-L., and C.G.C. approved final version of manuscript.