INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

THE CHRONIC MANIPULATION of vagal activity is difficult to achieve and as a result little is known about its consequences. The acute administration of morphine depresses cardiovascular function in many animal species. Decreases in heart rate, blood pressure, and cardiac output have been observed following epidural, intracardiac, and intravenous morphine administration in rats and dogs (13, 37, 38, 39). These effects can often be diminished or abolished by vagotomy (13, 38, 39), indicating that increased vagal activity is at least partially responsible for the reduced function. This "vagotonic" action of morphine is beneficial in situations when excess adrenergic activity threatens cardiovascular stability. For example, acute morphine reduced the vulnerability to ventricular fibrillation in psychologically stressed dogs (8, 9), an effect that was significantly diminished by vagotomy or atropine. In addition, morphine is often used to treat acute myocardial infarction, where its clinical utility derives in part from its ability to balance autonomic function through the reassertion of cardioprotective vagal influences.

The maintenance of normal vagal control of heart rate is important for long-term survival following acute myocardial infarction. Heart rate variability is often used as an index of the quality of vagal control. Although both the sympathetic and parasympathetic systems participate in the control of heart rate, faster parasympathetic components are primarily responsible for the instantaneous adjustments in heart rate that follow changes in respiration and posture. This beat-to-beat variation in heart rate, therefore, indicates the strength of vagal influences in controlling heart rate. Kleiger et al. (19) reported that survival rates following acute myocardial infarction were highly correlated with heart rate variability. Patients with higher heart rate variability (increased vagal control) 1 wk after myocardial infarction were five times more likely to survive 3 yr than patients with less variability.

Although the vagotonic actions of acute morphine have been well documented, few studies have examined the effects of chronic morphine administration and the resulting chronic vagal tone on the control of cardiovascular function. Leung et al. (25) found no differences between morphine-treated rats and controls when they evaluated vagally mediated bradycardia and hypotension. These findings differ from data obtained in dogs (33, 49). When naloxone was administered to dogs treated with morphine for 7 days, heart rate and blood pressure increased sharply (49). Heart rate remained elevated for 4 h despite the return of blood pressure to pretreatment control after 1 h. This persistent tachycardia in addicted, but normotensive, animals suggests that chronic morphine may have impaired the ability of the vagus to return heart rate to normal. We similarly found that vagal bradycardia was impaired in dogs implanted with subcutaneous morphine pellets for 7 days (33).

Disturbances in parasympathetic function may contribute to fatalities in opiate addicts. Lipski et al. (27) reported cardiac arrhythmias in more than 50% of otherwise healthy opiate addicts and suggested that the resulting cardiac vulnerability might contribute to narcotic-associated fatalities. The high incidence of bradyarrhythmias in this study may indicate a tonic increase in vagal activity, a loss of parasympathetic reserve, and, perhaps, the inability to react appropriately to changes in sympathetic influences. In addition, most fatalities classified as heroin overdoses occur among dependent, regular users, not novices, and often occur several hours after the last use. In the majority of cases, death occurs in the absence of toxic concentrations of the narcotic in question (7). These data collectively suggest that chronic adaptations rather than acute toxicity may contribute to heroin-related deaths.

Thus parasympathetic control of cardiovascular function appears to be altered in opiate addicts and related animal models. This study employed a canine model to test the hypothesis that persistent increases in efferent vagal traffic associated with chronic morphine treatment would reduce vagal reserve and attenuate or downregulate the parasympathetic control of heart rate. Parasympathetic and sympathetic contributions to heart rate variability were examined before, during, and after chronic morphine treatment with the aid of power spectral analysis. Autonomic blockade was utilized to determine whether the heart itself adapts to chronic morphine. Finally, direct vagal nerve stimulations were conducted to determine whether chronic morphine alters efferent bradycardic responses. The findings provide insight into basic autonomic adaptations to chronic vagal stimulation. The results are pertinent to tolerance and dependence models and are particularly important because a resurgence in opiate abuse has occurred in recent years (14). The study also provides relevant clinical information because opiates are increasingly prescribed for the relief of chronic pain (6).

	METHODS

Top Abstract Introduction Methods Results Discussion References

Dog Selection and Training

Catheter Implantation and Treatment Protocol

1

1

Serum morphine was determined by radioimmunoassay (Coat-A-Count; Diagnostic Products, Los Angeles CA).

Power Spectral Analysis Evaluation of Conscious Animals

Recording and analysis of heart rate signals. The heart rate signal was captured from a surface electrocardiogram (ECG) and continuously displayed on a monitor (Hewlett-Packard 78354A). The signal was relayed to a 12-bit analog-to-digital convertor (ADC, CIO-DAS16 Jr) connected to an IBM-compatible personal computer equipped with a 486-MHz microprocessor.

The system digitized heart rate signals at a rate of 512 Hz. The signals were then analyzed by an algorithm that allows for continuous, online PSA in real time (21). The digitized signals were truncated into 32-s time segments (windows). For each segment, the algorithm used fast Fourier transforms to estimate the power density of the spectral components. The components were quantified by integration of the area of power spectral density between specific upper and lower frequency limits. The selected frequency ranges were 1) high frequency (HF), 0.15-0.40 Hz; 2) low frequency (LF), 0.08-0.15 Hz; and 3) very low frequency (VLF), 0.01-0.08 Hz. The graphical results were continuously displayed on the monitor and printed. Numerical output was listed and printed at the end of each recording session. Data were recorded for later retrieval on an external Zip drive. All data were edited manually for artifacts caused by ectopic or premature beats, ECG signal interference, or gross subject movement. The raw power spectral data were routinely normalized within subjects using the standard formulas, HF/(HF + LF) and LF/(HF + LF). The formulas allow a determination of the relative influence of high- and low-frequency fluctuations on heart rate variability (minus the VLF).

Morphine initiation protocol. The heart rate power spectrum was observed and recorded during the first 3 h of morphine treatment on day 0 of the protocol. Heart rate was monitored continuously while the dog rested quietly in the sling. Every attempt was made to reduce extraneous noise and other disturbances in the laboratory. The room temperature was controlled at 24-26°C. When a steady state was achieved, a 10- to 20-min baseline recording was obtained during saline infusion. The morphine infusion was then initiated, and the power spectrum was recorded and printed continuously for 3 h. Blood samples for plasma morphine determination were obtained via an intravenous catheter at 15, 30, 60, 90, 120, 150, and 180 min. The PSA data were divided into time periods for analysis as follows: 0-15 min, 15-30 min, 30-60 min, 60-90 min, 90-120 min, 120-150 min, and 150-180 min. PSA data were also obtained from control animals for the same number of minutes during saline infusion.

Autonomic blockade protocol. Autonomic blockade with simultaneous PSA was conducted before treatment (day 4), and both early (day 2) and late (day 10) in the treatment protocol. A venous catheter was inserted into the forelimb to obtain blood samples and to administer drugs. Heart rate was monitored continuously while the dog rested quietly in the sling as described above. When a steady state was achieved, the heart rate power spectrum was recorded for 10-20 min. The muscarinic antagonist atropine methyl bromide (AMB, 75 µg/kg iv) was then administered to induce parasympathetic blockade. Because AMB does not readily cross the blood-brain barrier, the secondary central effects of atropine were avoided (12). The selected dose of AMB results in complete muscarinic blockade without interrupting ganglionic transmission (47). The adequacy of the blockade was verified in each animal by demonstrating no further change in heart rate after doubling of the dose. Sympathetic blockade with atenolol (3.0 mg/kg iv) was then induced to determine the heart rate in the absence of neural input (intrinsic heart rate). Blockade with atenolol was previously verified in animals similarly pretreated with atropine and challenged with isoproterenol (1 µg). The heart rate power spectrum was recorded continuously during the experiment. Data were edited as described above, and 5-10 min of steady-state data following each injection were used for analysis.

Plasma Catecholamines

Surgical Procedure

1

1

Vagal Stimulation

Heart rate was sampled during each 15-s stimulation. A 105-s interval followed each stimulation to allow heart rate to return to normal. The times to reach the minimum heart rate during vagal stimulation and to return to the prestimulation heart rate were recorded for each frequency. Data were sampled at appropriate times, digitized, and recorded online (MacLab).

To separate chronic adaptive responses from effects due to morphine circulating at the time of the experiment, the vagal stimulation protocol was also conducted in animals treated acutely with morphine (1 mg/kg sc). This dose produced plasma morphine concentrations of approximately double those observed in chronically treated animals (230 ± 14 vs. 108 ± 5 ng/ml). The dose was selected to produce central nervous system (CNS) morphine concentrations equal to those in animals treated with chronic morphine. After acute morphine administration, CNS morphine concentrations in dogs are approximately one-half those observed in plasma (29, 30). Surgical preparation was as described above for chronically treated animals.

Statistics

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Body Mass

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Weight (kg) of saline control and chronic morphine-treated dogs during the 14-day treatment period. Values are means ± SE; n = 12 for saline controls and 11 for chronic morphine. * Significantly different (P < 0.01) from day 0 and from controls on the same day.

Plasma Morphine

View larger version (11K): [in this window] [in a new window]   Fig. 2.   Plasma morphine concentrations (ng/ml) during first 3 h of morphine infusion and during the 14-day treatment period. Values are means ± SE; n = 10 or 11.

Initial Exposure to Morphine

View larger version (13K): [in this window] [in a new window]   Fig. 3.   Change in heart rate [beats/min (bpm)] during first 3 h of saline or morphine infusion. Values are means ± SE; n = 7 for saline controls and 10 for morphine-treated animals. * Significantly different (P < 0.01) from baseline; + significantly different (P < 0.05) from saline controls at the same time.

View larger version (12K): [in this window] [in a new window]   Fig. 4.   Linear regression analysis of the change in heart rate as a function of plasma morphine concentration (ng/ml) during the first 3 h of morphine infusion; r = 0.74 (P < 0.01).

PSA. PSA data from the initial 3-h saline or morphine period are presented in Fig. 5. Variation in the raw power spectral data necessitated that the data be normalized within subjects. Normalization of HF and LF power using the formulas HF/(HF + LF) and LF/(HF + LF), respectively, provides an estimate of parasympathetic-sympathetic balance. As illustrated (Fig. 5), HF power represented 70-90% of combined power (excluding VLF) while LF power represented the remaining 10-30%. This confirms the predominance of vagal control of heart rate variability in the conscious dog (36). Normalized HF and LF powers were similar in both groups, indicating that morphine infusion did not change relative parasympathetic or sympathetic contributions to heart rate variability. The stability of autonomic balance is also demonstrated by the LF-to-HF ratio, which did not change significantly during the initial 3 h in either group (not shown). The VLF power represents a small portion of the total power and did not change significantly in either group during the 3 h. The raw HF power declined during the 3-h period in morphine-treated animals and was significantly lower than baseline by 60 min (P < 0.05, data not shown).

View larger version (17K): [in this window] [in a new window]   Fig. 5.   High-frequency (HF) and low-frequency (LF) power spectral components of heart rate expressed in normalized units (nu) for the first 3 h of saline or morphine infusion. Values are means ± SE; n = 7 for saline controls and 10 for morphine-treated animals.

Early and Late Effects of Chronic Morphine

View larger version (24K): [in this window] [in a new window]   Fig. 6.   Heart rates at rest, after atropine methyl bromide (AMB), and after atenolol in dogs before treatment and on days 2 and 10 of saline or morphine infusion. Values are means; n = 8 and 10 for saline and morphine dogs, respectively. # Different from saline day 2; + different from saline day 2 and from before (Pre) morphine treatment and day 10; * different from saline on same day and from before morphine treatment; ** different from before morphine treatment. All labeled differences are significant at P < 0.05.

PSA. PSA data obtained before treatment and on days 2 and 10 of the saline or morphine infusion are presented in Table 3. Although the control and morphine-treated dogs differed initially, within-group analyses indicated no change from baseline in any PSA variable on day 2 or 10 of saline or morphine infusion. This is again best illustrated in the normalized units. A representative tracing is presented in Fig. 7. The administration of AMB nearly abolished the HF power, verifying that HF fluctuations in heart rate primarily reflect vagal influences. The LF power was substantially reduced by atropine, indicating that parasympathetic influences also contribute to heart rate variation in this range. As reflected by the normalized units, AMB shifted the autonomic balance toward greater sympathetic influence. Atenolol further reduced HF power and almost completely abolished the LF power. Relative parasympathetic and sympathetic control of remaining heart rate variability was approximately equal following atenolol, as reflected by the HF (52%) and LF (48%) normalized units. The HF and LF power after autonomic blockade were unaffected by morphine on days 2 and 10. The VLF power was reduced by the administration of AMB and almost eradicated with subsequent administration of atenolol. There were no differences in the VLF within treatment groups over time or between treatment groups before autonomic blockade. A small increase in VLF was observed in morphine-treated animals after blockade; however, any conclusion in this regard is suspect because VLF power was very near the baseline noise in all blocked animals.

View larger version (31K): [in this window] [in a new window]   Fig. 7.   Representative tracing of continuous, online, real-time power spectral analysis of heart rate (HR) variability during sequential autonomic blockade. ECG, electrocardiogram; HPSD, heart rate power spectral density; HHF, heart rate high frequency; HLF, heart rate low frequency; HVLF, heart rate very low frequency.

Plasma Catecholamines

View larger version (15K): [in this window] [in a new window]   Fig. 8.   Plasma catecholamines in saline control and morphine-treated dogs pretreatment and on days 2 and 10. Values are means ± SE; n = 12 for both groups. * Significantly different from pretreatment (P < 0.05).

Vagal Stimulation Data

The right vagus nerve was stimulated at 0.5, 1.0, 2.0, and 4.0 Hz. Both vagi were ligated before stimulation to eliminate afferent vagal traffic while efferent bradycardic response was obtained. In all groups, heart rate declined further with each increase in the frequency of stimulation, reaching a decrease of 55-60 beats/min at 4 Hz. The frequency-response relationship was not significantly different between the three groups at any frequency.

The kinetics of heart rate responses during vagal stimulation were monitored, and the times required for heart rate to reach 50% of the maximum decrease during stimulation and to return to 50% of the prestimulation rate are presented in Figs. 9 and 10, respectively. In all groups, the heart rate declined more quickly as the frequency of vagal stimulation increased (Fig. 9). The 50% down-time intervals were significantly different among the three groups across frequencies (P < 0.01), with the control group reaching 50% of maximum bradycardia more quickly, followed by the chronic and acute morphine groups. In all groups, heart rate returned to the prestimulation rate more quickly as the frequency of stimulation increased (Fig. 10). Although chronic morphine appeared to slow the recovery, these time intervals were not significantly different among treatment groups at any frequency.

View larger version (27K): [in this window] [in a new window]   Fig. 9.   Time (s) for heart rate to reach 50% maximum bradycardia during right vagal nerve stimulation for controls and for dogs treated with acute or chronic morphine. Values are means ± SE; n = 10 for controls, 11 for chronic morphine, and 7 for acute morphine. Times were significantly different between the 3 groups across frequencies (P < 0.01).

View larger version (24K): [in this window] [in a new window]   Fig. 10.   Time (s) for heart rate to return to 50% prestimulation heart rate on termination of right vagal nerve stimulation for controls and for dogs treated with acute or chronic morphine. Values are means ± SE; n = 10 for controls, 11 for chronic morphine, and 7 for acute morphine.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The experiments presented above were conducted to determine whether chronic morphine produced a persistent increase in efferent vagal outflow and a resultant reduction in parasympathetic control of the heart. Dogs received subcutaneous morphine for 2 wk in doses to maintain plasma morphine concentrations between 80 and 120 ng/ml. This concentration of morphine is sufficient to induce physical dependence in dogs. Yoshimura et al. (49) implanted miniosmotic pumps subcutaneously in dogs and delivered morphine to produce plasma concentrations of 40-50 ng/ml. After 8 days of treatment, the administration of the opiate antagonist naloxone (1 mg/kg) produced both behavioral and physiological signs of withdrawal, including hyperactivity, biting, tremors, nausea, tachycardia, and changes in the electroencephalogram. The rise in blood pressure following naloxone, an index of opiate dependence, was 45-70 mmHg in their dogs. These results are consistent with those we previously obtained in dogs after 7 days of subcutaneous morphine, which produced plasma concentrations of 100 ng/ml. In those animals, naloxone administration increased blood pressure by 30 mmHg (33).

Morphine reduces heart rate through activation of vagal mechanisms. Pur-Shahriari et al. (37) demonstrated a decline in heart rate in dogs following right atrial injection of morphine. Hotvedt and Refsum (13) provided evidence that morphine-induced bradycardia arises from an increase in vagal activity. They demonstrated that the reduction in heart rate following thoracic epidural morphine administration in dogs was reversed by bilateral cervical vagotomy. Similar results have been demonstrated in rats following intravenous morphine administration (38, 39). The decline in heart rate observed during the first 3 h of morphine infusion in the present study is consistent with this bradycardic effect of acute morphine.

Because morphine stimulates vagal outflow, we predicted that HF fluctuations in heart rate would increase with morphine treatment. However, the power of the HF component represents fluctuations in heart rate mediated by efferent vagal activity and not vagal tone per se (28). If morphine produces an increased but invariant or less variant vagal outflow, fluctuations in heart rate could decrease. Lee et al. (23) measured the effects of the opioid anesthetic fentanyl on the arterial pressure power spectra in anesthetized rats. The power spectral components associated with fluctuations in arterial pressure reflect activity of the same control mechanisms as the heart rate power spectral peaks (36). Fentanyl substantially reduced HF power in these animals, which may also reflect an invariant or less variant increase in vagal activity. The decrease in HF power during the initial 3 h of morphine treatment in the present study is consistent with this theory. We have also observed this decline in HF power in ongoing studies in dogs given higher doses of acute morphine.

Sequential autonomic blockade provided several indexes of parasympathetic and sympathetic activity. The heart rate following parasympathetic blockade with AMB represents the combined influence of intrinsic heart rate and ambient sympathetic tone. The difference between resting heart rate and this post-AMB heart rate has been described as the vagal effect (35) because it reflects the amount of vagal activity that was required to maintain the lower resting heart rate before atropine. Subsequent administration of atenolol produced complete autonomic blockade, revealing the intrinsic heart rate. The difference between the post-AMB heart rate and intrinsic heart rate is a measure of ambient sympathetic tone.

The vagal influence on resting heart rate and the ambient sympathetic tone were increased on days 2 and 10 in morphine-treated animals. The increase in sympathetic tone suggests a compensatory sympathetic response to the increase in vagal activity induced by morphine treatment. This increase in sympathetic tone may have been responsible for returning the resting heart rate to normal on day 10. The intrinsic heart rate on day 2 was depressed to the same degree as the resting heart rate, suggesting that the decline in intrinsic heart rate was responsible for the change in resting heart rate. However, because sympathetic tone was also greater, an increase in parasympathetic activity must have accompanied this enhanced ambient sympathetic activity. The intrinsic heart rate remained depressed on day 10. The lower intrinsic heart rates on days 2 and 10 are not due to the acute influence of morphine because intrinsic rate was not lower in animals following acute morphine administration.

A decline in intrinsic heart rate may reflect a fundamental change in the function of sinoatrial nodal cells. Several possibilities have been suggested in this regard, including alterations in mechanical factors (26), cellular metabolism (17), and potassium kinetics (15). The intrinsic heart rate following double blockade may also involve contributions from other chronotropic agents. The neuropeptide vasoactive intestinal peptide (VIP) is coreleased with acetylcholine from postganglionic parasympathetic neurons and produces an opposing increase in heart rate (40). This "excess tachycardia" is demonstrated by an additional reduction in heart rate after ganglionic blockade or vagotomy subsequent to the usual double receptor blockade (4, 41, 44). Thus the reduction in the intrinsic heart rate suggests that chronic morphine may reduce ganglionic transmission, the synthesis of VIP, or the release of VIP from the parasympathetic nerve terminals, thereby reducing the excess tachycardia and lowering the apparent intrinsic heart rate. Alternatively, the chronic parasympathetic activation associated with morphine treatment may deplete the neurons of their stores of VIP.

The intrinsic heart rates obtained in control animals (185 beats/min) in the present study were somewhat higher than those reported for dogs elsewhere in the literature (10) but virtually identical to those reported in the initial studies of excess tachycardia (41). Lower intrinsic heart rates in some studies may reflect a reduction in the excess tachycardia described above due to the gangliolytic properties of higher doses of atropine. The dose of AMB selected for our study was specifically titrated to provide for complete muscarinic blockade without blocking the ganglion.

An increase in plasma catecholamines paralleled the augmentation of ambient sympathetic activity after 2 days of morphine treatment. The sympathetic activity remained high through day 10 despite the return of plasma catecholamines toward the baseline concentrations. This suggests an increase in the responsiveness to catecholamines. This normalization of circulating catecholamines would be consistent with observations by Leung et al. (24), who found no change in arterial catecholamines in rats following 3 wk of morphine. Plasma concentrations were not determined during their treatment protocol, which precludes comparisons with our early observations of elevated catecholamines.

Contrary to expectations, efferent vagal control of the heart during direct vagal nerve stimulations was not altered by chronic morphine treatment. This contrasts with our earlier observation, when we found a significant attenuation of efferent vagal bradycardia in dogs treated with subcutaneous morphine pellets for 7 days (33). These conflicting results were not due to different plasma morphine concentrations because the concentrations were similar throughout the two protocols. The animals may have become tolerant to the effects of morphine between days 7 and 14. Leung et al. (25) similarly found no effect of chronic morphine on vagal bradycardia in rats that had received morphine in drinking water for 3 wk. The administered dose of morphine had been increased in their study but only during the first 8 days. Thereafter, the dose remained constant. Progressively increasing the dose of morphine might reduce the likelihood that animals would accommodate to the effects of the drug.

Kosterlitz and Taylor (20) first reported the ability of morphine to inhibit vagal bradycardia in anesthetized rats and rabbits but were unable to show this effect in guinea pigs. Acute morphine has been shown to reduce the bradycardic response to vagal nerve stimulation (48) and to stimulation of cholinergic fibers in the sinoatrial node (18) in isolated rabbit heart preparations. The reduced vagal function was attributed to morphine-mediated inhibition of acetylcholine release from postganglionic nerve fibers (18). Musha et al. (32) found that acute morphine in doses similar to those used in the present study did not alter vagal bradycardia in anesthetized dogs. Similarly, acute morphine had no effect on stimulated vagal bradycardia in the present study. The 1 mg/kg dose of morphine administered in our acute experiments produced a plasma concentration of approximately double that of chronic animals. The acute dosage was selected to ensure that the resulting central interstitial concentrations would equal or exceed those obtained during chronic administration (29, 30). Although central morphine concentrations lag behind those in the plasma after acute administration, we made the assumption that they would eventually equilibrate with and approach those in the plasma during continuous delivery. These conflicting effects of acute morphine cited above may be due to the different experimental preparations (isolated versus intact) or to species differences in the distribution of µ-receptors.

Neither acute nor chronic morphine treatment had any effect on the absolute decrease in heart rate observed at each vagal stimulation frequency. Perhaps not unexpectedly, the rate of decline and rate of recovery of heart rate during and after vagal stimulation was proportional to the frequency of stimulation. The time to reach 50% of each decline in heart rate was longer in dogs that had received morphine chronically. This time was extended even further in animals treated with morphine acutely. These results suggest that acute morphine altered the kinetics of vagal bradycardia potentially by decreasing the rate of acetylcholine release or by increasing its rate of degradation. Slower vagal responses might render the heart more vulnerable to sympathetically mediated arrhythmias. Acute morphine inhibits acetylcholine release in a number of systems (3, 11, 43, 45), but the effects of chronic morphine treatment are less consistent. Some studies have found no change in release (5, 31), whereas others have reported a decrease in release with chronic morphine (16, 22). However, none of these studies measured the rate of acetylcholine release. Because the rate of recovery after the stimulus was terminated was unaltered, an increase in acetylcholine degradation seems less likely. In fact, acute morphine inhibits serum and tissue cholinesterase in vitro, albeit at very high morphine concentrations (46). A chronic effect of morphine on myocardial cholinesterase has not been determined.

In summary, we predicted that chronic morphine accompanied by persistent efferent vagal traffic would attenuate or downregulate parasympathetic control of myocardial function on the basis of preliminary observations made after 1 wk of treatment with morphine. The initial decrease in heart rate and HF power spectrum are consistent with increased vagal activity with morphine. However, after 2 wk of treatment, an increasing vagal effect, a return of the resting heart rate to normal, and a progressive increase in ambient sympathetic activity all suggest that subsequent sympathetic and parasympathetic compensations may have overridden any early vagal impairment. Collectively, these results suggest an adaptive, compensatory response in sympathetic nervous activity to chronic morphine or to the resultant vagal outflow. The lower intrinsic heart rates with chronic morphine may reflect fundamental changes in sinoatrial nodal cell function or, alternatively, an attenuation in nonadrenergic, noncholinergic (VIP) mechanisms. The altered kinetics of vagal bradycardia suggests that morphine inhibits acetylcholine release. Thus, if morphine did decrease parasympathetic reserve through downregulation early in the treatment, the effects were most likely subsequently masked by greater yet sympathetic adjustments. Muscarinic receptors and associated inhibitory G proteins were upregulated in these same animals (L. D. Napier, S. C. Roerig, D. A. Yoshishige, B. A. Barron, and J. L. Caffrey, unpublished observations), suggesting a secondary parasympathetic response to the augmented sympathetic activity indicated in the present study.

The clinical consequences of these findings are not entirely clear. However, the progressive compensatory increase in sympathetic activity during exposure to morphine might have significant cardiovascular consequences if the pharmacologically augmented vagal opposition were suddenly withdrawn. The unopposed sympathetic activity during abrupt withdrawal might raise clinical concerns, especially for patients with existing arrhythmias and/or underlying ischemic heart disease. The resulting sympathetic imbalance may explain the common observation that opiate-related deaths are often associated with nontoxic opiate concentrations (7).