Doxapram Only Slightly Reduces the Shivering Threshold in Healthy Volunteers

Abstract

We determined the effects of doxapram on the major autonomic thermoregulatory responses in humans. Nine healthy volunteers were studied on two days: Control and Doxapram (intravenous infusion to a plasma concentration of 2.4 ±0.8 μg/mL, 2.5 ±0.9 μg/mL, and 2.6 ±1.1 μg/mL at the sweating, vasoconstriction, and shivering thresholds, respectively). Each day, skin and core temperatures were increased to provoke sweating, then reduced to elicit peripheral vasoconstriction and shivering. We determined the sweating, vasoconstriction, and shivering thresholds with compensation for changes in skin temperature. Data were analyzed with paired t tests and presented as means ± SDs; P < 0.05 was considered statistically significant. Doxapram did not change the sweating (Control: 37.5±0.4°C, Doxapram: 37.3±0.4°C, P=0.290) or the vasoconstriction threshold (36.8±0.7 vs. 36.4±0.5°C; P=0.110). However, it significantly reduced the shivering threshold from 36.2±0.5 to 35.7±0.7°C (P=0.012). No sedation or symptoms of panic were observed on either study day. The observed reduction in the shivering threshold explains the drug's efficacy for treatment of postoperative shivering; however, a reduction of only 0.5°C is unlikely to markedly facilitate induction of therapeutic hypothermia as a sole agent.

Introduction

Overwhelming evidence in animals indicates that even mild hypothermia provides substantial protection against cerebral (1) and myocardial ischemia (2). Mild hypothermia has been shown to improve outcome after cardiac arrest in humans (3). Many of these studies target core temperatures between 33° and 34°C.

Induction of therapeutic hypothermia in patients having acute myocardial infarction or stroke may be compromised because tiny reductions in core temperature trigger aggressive thermoregulatory defenses (4). The major autonomic thermoregulatory responses in humans are sweating, vasoconstriction, and shivering. Each response is characterized by a threshold, which is defined as the triggering core temperature.

Dopamine is among the most important thermoregulatory neurotransmitters, and it is well established that increasing preoptic dopamine concentrations provokes hypothermia in mammals (5). Although doxapram stimulates release of dopamine from carotid bodies in rats (6), it is known to have central effects (7) that are probably, at least in part, similarly mediated. As might thus be expected, doxapram is a useful treatment for postanesthetic shivering (8). Doxapram produces a dose-dependent and substantial reduction in the shivering threshold in rabbits (9). The magnitude of this inhibition, if similar in humans, would be clinically important. Equally important is whether doxapram reduces the shivering threshold from its normal value near 35.5°C (10) to ≈34°C which may provide protection against cerebral or myocardial ischemia. We thus tested the hypotheses that doxapram comparably reduces the sweating, vasoconstriction, and shivering thresholds — and that the reduction is clinical important (i.e., ≈1°C).

Methods

With approval of the Human Studies Committee at the University of Louisville and written informed consent, we studied nine healthy volunteers (five men and four women). None was obese, taking medication, or had a history of thyroid disease, dysautonomia, or Raynaud's syndrome.

Protocol

The volunteers fasted 8 hours before each study day. They dressed minimally and rested supine on a standard operating room table. Ambient temperature was maintained near 21°C. Each volunteer was studied on two randomly assigned days: 1) Control — no drug — and 2) doxapram hydrochloride (A. H. Robins, Inc.) at a target plasma concentration of 4 μg/mL.

Doxapram was infused using a modification of the simplified infusion regimen of Clements et al. (11), which was stated to produce a constant plasma concentration of 2 μg/mL. In this case with an objective of 4 μg/mL, the infusion rates were doubled under the assumption of linear pharmacokinetics. Specifically, doxapram was infused at a rate of 6 mg.min^-1.70 kg^-1 for the first 15 minutes; at a rate of 4 mg.min^-1.70 kg^-1 for the next 15 minutes; at a rate of 3 mg.min^-1.70 kg^-1 for the next 30 minutes; at a rate of 2 mg.min^-1.70 kg^-1 for the next hour, and subsequently, at a rate of 1 mg.min^-1.70 kg^-1. The infusion was started 15 minutes before starting thermal manipulation to allow establishment of steady-state plasma drug concentration; the infusion was continued until shivering was detected.

Sweating, Vasoconstriction, & Shivering Thresholds

Core temperature was recorded from the tympanic membrane using Mon-a-therm thermocouples (Mallinckrodt Anesthesiology Products, Inc., St. Louis, MO). Mean skin-surface temperature and cutaneous heat transfer were calculated from measurements at 15 area-weighted sites. Temperatures were recorded at one-min intervals from thermocouples connected to calibrated Iso-Thermex® thermometers having an accuracy of 0.1°C and a precision of 0.01°C (Columbus Instruments, Corp., Columbus, OH).

Sweating was continuously quantified on the left upper chest using a ventilated capsule (4). We considered a sweating rate > 40 g.m^-2.h^-1 for at least 5 minutes to be significant (12). Absolute right middle fingertip blood flow was quantified using venous-occlusion volume plethysmography at 5-minute intervals (13). The vasoconstriction threshold was determined post hoc by an observer blinded to treatment and core temperature.

As in previous similar studies (12), we used systemic oxygen consumption to quantify shivering. A DeltaTrac metabolic monitor (SensorMedics Corp., Yorba Linda, CA) was used in canopy mode. Initiation of shivering threshold was determined post hoc by an observer blinded to treatment and core temperature. End-tidal PCO₂ was sampled from a catheter inserted into one nostril; gas removed from the catheter was returned to the canopy of the metabolic monitor.

Heart rate, end-tidal PCO₂, and oxyhemoglobin saturation (SpO₂) were measured continuously using pulse oximetry, and blood pressure was determined oscillometrically at 5-minute intervals at the left ankle.

Sedation was evaluated by using the responsiveness component of the Observer's Assessment of Alertness/Sedation (OAA/S) score (Table 1) (14) at several times: before starting drug administration, before thermal manipulation started, and at each threshold. An investigator blinded to core temperature and treatment evaluated sedation. Blood for doxapram analysis was sampled from the central catheter before drug administration (blank sample) and at each threshold. Blood was centrifuged, and the plasma was removed and stored at –40°C for up to 2 months. The plasma samples were then transferred to another freezer and stored for up to 1 week until analysis. Doxapram and its metabolites are stable when stored ≤ -20°C.

Table 1.

Responsive Component of the Observer's Assessment of Alertness/Sedation Scale

Score	Responsiveness
5	Responds readily to name spoken in normal tone
4	Lethargic response to name spoken in normal tone
3	Responds only after name is spoken loudly and/or repeatedly
2	Responds only after mild prodding or shaking
1	Dose not respond to mild prodding or shaking

Doxapram concentrations were determined in triplicate by gas chromatography/mass spectrometry using the technique of selected ion monitoring (GC/MS-SIM). Reference doxapram hydrochloride was obtained from USP (U.S. Pharmacopeia, Rockville, MD). The method used solid phase extraction columns (Strata-X, Phenomenex, Torrance CA) to extract doxapram from plasma along with internal standard (diazepam) to compensate the drug loss during extraction and GC/MS processes. Doxapram and diazepam were separated by gas chromatography according to their different GC retention times (7.0, 5.4 min, respectively) in a capillary column (DB5, Agilent, Palo Alto, CA) that was temperature programmed from 150° to 290°C at 32°/min. The drugs were then detected by MS using SIM of characteristic ions (doxapram m/z 100, 378; diazepam m/z 256, 283) following electron ionization. Quantification used standard responses measured by ion peak area ratios versus amounts of analyte/internal standard. Responses were linear across the doxapram plasma concentration range of 0.25 – 5 μg/mL. Limits of detection and quantification were less than 0.25 μg/mL. Sample replicates were assayed on different occasions. The intra-assay coefficient of variation (CV) was 8.8%, while inter-assay CV was 21%.

Nausea and vomiting incidence and severity were recorded by blinded observers before commencement of drug administration, before thermal manipulation started, and at each threshold during each study. Nausea and vomiting incidence and severity were reported by the volunteer using a four-point scale: 1 = none, 2 = mild, 3 = moderate, or 4 = severe.

Because doxapram is a potent agent to cause panic (15), we scored the severity of panic attack symptoms by presence or absence of seven selected items of DSM-IV-TR diagnostic criteria for Panic Attack (Table 2) (16). These were assessed at the same times as the OAA/S score. The total number of the symptoms present at each time was used for statistical analysis.

Table 2.

Panic Symptoms.

Score	Symptom
1	Nausea or abdominal distress
2	Paresthesia (numbness or tingling sensations)
3	Feeling of choking
4	Palpitations, pounding heart, or accelerated heart rate
5	Chest pain or discomfort
6	Derealization (feelings of unreality) or depersonalization (being detached from oneself)
7	Feeling dizzy, unsteady, light-headed, or faint

Data Analysis

The cutaneous contribution to the thermoregulatory responses — sweating, (17), vasoconstriction, and shivering — is linear (18). We thus used measured skin and core temperatures in °C at each threshold to calculate the core-temperature threshold that would have been observed had skin been maintained at a single designated temperature. For this purpose, we used Eq. 1 that corrects core temperature for cutaneous temperature, providing response thresholds that would have been observed at a designated core temperature:

$$T_{\text{core}\left(\text{calculated}\right)}=T_{\text{core}}+\left(\frac{\beta }{1-\beta }\right)[T_{\text{skin}}-T_{\text{skin}\left(\text{designated}\right)}]$$ 1

We have previously described the derivation and validation of this equation (12). We used a β of 0.1 for sweating (17) and a β of 0.2 for vasoconstriction and shivering (18). The designated skin temperature was set at 34°C, a typical intraoperative value.

We determined that a sample size of 9 would detect a 0.6°C difference in shivering thresholds between control and drug days with 89% power. Based on a similar study conducted with the same design in our laboratory with another drug (unpublished observation), we assumed that the difference in shivering thresholds on the two study days would have a standard deviation of 0.495 and correlation of 0.55.

Oxygen consumption and calculated respiratory quotient (RQ) were averaged during baseline (before commencement of drug infusion), during the cooling phase before shivering, and during shivering, respectively. The averaged values were then compared with two-way ANOVA (two factors; drug, period) and post-hoc Bonferroni/Dunn tests.

After confirming that sedation, nausea severity, and panic symptoms were similar at baseline on each study day, the difference in the scores at each measurement time (baseline, prewarming, sweating threshold, vasoconstriction threshold, and shivering threshold) between the drug day and the control day within each volunteer were compared with Kruskal Wallis tests. The incidence of nausea and combined nausea and vomiting was compared by Chi-square analyses. For the combined nausea and vomiting analysis, any score for nausea or vomiting exceeding “none” was considered positive.

Ambient temperature, humidity, mean arterial pressure, heart rate, SpO₂, and end-tidal PCO₂ on each study day were averaged within each volunteer across the warming and cooling periods; the resulting values were then averaged among volunteers. Results for each study day were compared using paired t tests. Plasma concentrations of doxapram at each threshold were compared using one-way ANOVA. All results are presented as means± SD or range, as appropriate; P < 0.05 was considered statistically significant.

Results

Five of the participants were male; four were female. They were 26 ± 5 years old, weighed 72 ± 13 kg, and were 175 ± 10 cm tall. Of the possible confounding factors that might influence thermoregulatory thresholds, ambient temperature, humidity, and heart rate were similar on each of the study days during the warming and cooling periods. In addition, End-tidal PCO₂ was significantly reduced, whereas SpO₂ and mean arterial pressure were both increased significantly by doxapram during the warming and cooling periods (Table 3).

Table 3.

Potential confounding factors.

	Control Day	Doxapram Day	P
Values were recorded at 5-minute intervals and averaged over the sweating-to-shivering period.
Data are presented as means ± SDs; results were compared with paired t-tests.
Mean Arterial Blood Pressure (mmHg)	89 ± 10	102 ± 6	0.005
Heart Rate (beats per min)	71 ± 7	76 ± 7	0.065
End Tidal PCO2 (mm Hg)	38 ± 2	36 ± 2	0.023
SpO2 (%)	97 ± 1	98 ± 1	<0.001
Ambient temperature (°C)	23.0 ± 1.4	23.4 ± 2.2	0.400
Relative Humidity (%)	28.1 ± 11.4	21.8 ± 9.4	0.270

Oxygen consumption was not significantly influenced by doxapram (P=0.293), and it was similar between baseline and cooling phases. But it was significantly increased during shivering compared to both baseline and the cooling phase (P<0.001). RQ was not significantly influenced by doxapram (P=0.195), nor by the periods of the study (P=0.458) (Table 4).

Table 4.

Oxygen consumption and respiratory quotient on the control and doxapram study days.

	Control Day	Doxapram Day
Values were recorded at 1-minute intervals and averaged in each period. Data are presented as mean ± SD. There were no statistically significant differences between the study days for any period.
Oxygen consumption (mL/min)
Baseline period	217 ± 43	210 ± 16
Cooling period	226 ± 58	250 ± 53
Shivering period	315 ± 89	351 ± 51
Respiratory Quotient (RQ)
Baseline period	0.84 ± 0.05	0.80 ± 0.05
Cooling period	0.84 ± 0.08	0.82 ± 0.04
Shivering period	0.82 ± 0.05	0.79 ± 0.03

None of the study participants was sedated at baseline on either study day (OAA/S score of 5 for all volunteers), and no sedation was observed during the study. None of the volunteers was nauseated at the start of the study (nausea severity score of 1) on either study day. And, as with sedation, no significant nausea was observed during the study. None of the volunteers experienced panic at baseline on either study day (panic symptom score total of 0). Panic symptoms during doxapram infusion were not significantly different from those of the control day (P = 0.06).

On the doxapram day, plasma concentrations of doxapram were similar at the sweating, vasoconstriction, and shivering thresholds with plasma concentrations of 2.4 ± 0.8 [1.4-3.5] μg/mL, 2.5 ± 0.9 [1.4-3.6] μg/mL, and 2.6 ± 1.1 [1.3-4.8] μg/mL; (mean ± SD [range]) for each threshold, respectively (P = 0.930). The sweating thresholds were similar on the doxapram (37.3 ± 0.4°C) and control (37.5 ± 0.4°C) days (P = 0.290; Table 5). The vasoconstriction thresholds were also similar on the doxapram (36.4 ± 0.5°C; Table 5) and on the control (36.8 ± 0.7°C) day (P = 0.110). In contrast, doxapram reduced the shivering threshold by 0.5 ± 0.4°C from 36.2 ± 0.5°C on the control day to 35.7 ± 0.7°C on the doxapram day (P = 0.012, Fig. 1).

Table 5.

Mean Skin Temperatures, Core Temperatures, and Calculated Thermoregulatory Thresholds

	Control Day	Doxapram Day
Thresholds were calculated based on a designated mean-skin temperature of 34°C. Results are presented as mean ± SD and range for plasma doxapram concentrations. Only the thresholds were statistically compared, and only the shivering threshold differed significantly between the two treatment groups (P = 0.012).
Sweating
Mean Skin (°C)	36.6 ± 0.3	36.4 ± 1.0
Core (°C)	37.2 ± 0.3	37.0 ± 0.2
Threshold (°C)	37.5 ± 0.4	37.3 ± 0.4
Doxapram Concentration (μg/mL)	–	2.4 ± 0.8 [1.39-3.49]
Vasoconstriction
Mean Skin (°C)	32.9 ± 2.4	32.2 ± 1.5
Core (°C)	37.1 ± 0.3	36.8 ± 0.3
Threshold (°C)	36.8 ± 0.7	36.4 ± 0.5
Doxapram Concentration (μg/mL)	–	2.5 ± 0.9 [1.36-3.61]
Shivering
Mean Skin (°C)	30.0 ± 1.5	29.2 ± 1.3
Core (°C)	37.2 ± 0.3	36.9 ± 0.4
Threshold (°C)	36.2 ± 0.5	35.7 ± 0.7
Doxapram Concentration (μg/mL)	–	2.6 ± 1.1 [1.34-4.82]

Fig. 1.

Shivering thresholds in nine healthy volunteers. The open circles show the shivering threshold for each volunteer on the Control and Doxapram days; the filled squares are the group means (± SD). The shivering threshold was 0.5°C greater on the Control day than on the Doxapram day, P = 0.012.

Discussion

Although doxapram significantly reduced the shivering threshold, the reduction was only 0.5°C. In comparison, clonidine, at a dose of 75 μg, decreases the shivering threshold to the same extent as doxpram at the dose used in our study (19) and tramadol, at a dose of 250 μg, decreases the shivering threshold by 0.9°C, which is slightly greater than the change we obtained with doxapram (20). Both of those drugs have proven effective treatments for postoperative shivering. However, a reduction in the shivering threshold of this extent is insufficient to facilitate induction of therapeutic hypothermia. We thus conclude that doxapram as a sole agent will not serve for this purpose.

Doxapram increased SpO₂ (98 ± 1% versus 97 ± 1%) and reduced end-tidal PCO₂ (36±2 mmHg versus 38±2 mmHg), presumably as a consequence of doxapram's well-known ability to improve tidal volume and augment respiratory rate (21). The greater mean arterial pressure on the doxapram day (102 ± 6 mmHg versus 89 ± 10 mmHg) doubtless resulted from the established pressor effect of doxapram (21). Despite being statistical significant, these small respiratory and hemodynamic effects, which were not accompanied by increases in metabolic rates were of little clinical consequence and unlikely to cause adverse effects even in victims of stroke or myocardial infarction.

A target plasma concentration of 4 μg/mL was chosen as it appears to be the maximum therapeutic concentration that does not evoke serious side effects. A slightly greater plasma concentration of > 5 μg/mL is known to increase the risk of side effects (22). Our measured plasma concentration was within the range (1.5-3.0 μg/mL) that effectively increases the minute ventilation (23).

Although we used a published infusion protocol, the average plasma concentration at each threshold was considerably less than the targeted concentration of 4 μg/mL in all but one patient. We are unable to determine in the present study whether the differences between target and actual plasma levels resulted from differences in drug disposition, clearance, or nonlinear pharmacokinetic effects due to dosage scaling. Assuming a steady state plasma concentration was achieved before shivering occurred, the total plasma clearance calculated from our data was 12.9 ± 5 mL·kg^-1·min^-1 [range 5.8-21.5 mL·kg^-1·min^-1], a value larger than that reported by Clements et al. (11). However, Jamali et al. (24) found that 4 of 17 neonates had much greater clearance rates of doxapram (19.0-29.2 mL·kg^-1·min^-1) than the others. They suggested that the disposition kinetics of doxapram has a binominal distribution. If any of our participants were rapid metabolizers of doxapram, their elevated clearance rates could have profoundly reduced the actual mean plasma concentration of the group.

In summary, doxapram at a measured plasma concentration of ≈2.5 μg/mL reduced the shivering threshold significantly, but only by 0.5°C. This reduction explains the drug's efficacy for treatment of postoperative shivering. However, a reduction of only 0.5°C is unlikely to markedly facilitate induction of therapeutic hypothermia as a sole agent.