The Diagnosis and Treatment of Baroreflex Failure

Table 1. Table 1. Characteristics of 11 Patients with Baroreflex Failure.

Table 2. Table 2. Cardiovascular Characteristics of 13 Patients with Baroreflex Failure.

Among approximately 500 patients referred to the Autonomic Dysfunction Center at Vanderbilt University because of autonomic or blood-pressure problems, 11 patients with arterial baroreflex failure were identified (Table 1). The reasons for referral included evaluation for essential hypertension, suspected pheochromocytoma, uncontrolled hypertension, or recognition that the integrity of glossopharyngeal or vagal nerves had been compromised. Baroreflex failure was documented by the inability of infusions of pressor and depressor drugs to cause reflex bradycardia and tachycardia, respectively, in patients in whom wide and parallel variations in heart rate and blood pressure occurred in response to endogenous factors such as sedation and stress (Table 2). The diagnosis was supported in most patients by historical, physical, and physiologic information.

Baroreflex function was determined at rest in supine patients during hospitalization by measuring heart-rate responses to stepwise bolus injections of phenylephrine until a dose was found that raised systolic blood pressure by 25 mm Hg or more; the estimated fall in heart rate for an increase of 25 mm Hg in pressure was used as an index of the baroreflex control of heart rate. Similar studies were carried out with bolus injections of nitroprusside until a dose was found that lowered systolic blood pressure by 25 mm Hg. Changes in heart rate occurred within three minutes of drug administration. In four patients, the baroreflex effect on heart rate was determined from the Valsalva maneuver rather than by the administration of vasoactive drugs.

Causes of the baroreflex abnormalities in these patients are shown in Table 1. One patient had undergone surgical section of a glossopharyngeal nerve for intractable glossopharyngeal neuralgia after a neck injury that had apparently damaged the contralateral glossopharyngeal and vagal nerves. Three patients had undergone surgery and radiation therapy for carcinoma of the pharynx. Four patients had the familial paraganglioma syndrome, a genetic disorder characterized by multiple benign non-catecholamine-producing tumors of the carotid body, glomus jugulare, and glomus vagale¹⁸; these tumors damage the glossopharyngeal and vagal nerves. One patient had marked cell loss in the nuclei of each solitary tract of the brain stem caused by a degenerative neurologic disorder involving the brain stem and higher structures of the central nervous system; the cause was known from the results of an autopsy performed after the patient's death from pneumonia approximately one year after his evaluation. Finally, in two patients, no cause of the baroreflex failure could be identified.

After diagnosis, we monitored variations in blood pressure, heart rate, and plasma catecholamine concentrations in each patient¹⁹. Blood pressure was measured every four hours while the patients were supine and upright and during symptomatic episodes. Plasma catecholamine concentrations were measured while the patients were supine and upright and during episodes of symptoms.

Cold pressor tests were executed by having supine subjects put their right hands in a basin filled with half ice and half water for 60 seconds. Blood pressure and heart rate were measured before and at the end of this maneuver²⁰. Mental-arithmetic tests were performed by having supine subjects perform serial subtractions of 7 beginning with the number 200. Blood pressure and heart rate were determined before and 60 seconds after the start of the test. The isometric handgrip test was conducted as previously described²⁰.

Propranolol, given in a dose of 1 mg per minute intravenously for 10 minutes, and atropine, given as a bolus dose of 0.04 mg per kilogram of body weight intravenously, were administered to assess the sympathetic (beta₁-adrenoreceptor) and parasympathetic components of heart-rate control. Clonidine was given orally at a dose of 0.1 mg, and blood pressure and heart rate were monitored at 30-minute intervals during the succeeding 3 hours. The two-hour time point was used to determine to what degree clonidine reduced sympathetic activity, as reflected by a change in blood pressure and plasma norepinephrine concentration¹⁷.

Twelve normal subjects were studied in a similar fashion. In addition, in 8 patients with essential hypertension and 12 patients with autonomic failure due to the Bradbury-Eggleston syndrome, plasma catecholamine concentrations were measured and tests of responsiveness to clonidine and phenylephrine and the cold pressor test were administered. The Bradbury-Eggleston syndrome (idiopathic orthostatic hypotension) is a degenerative disorder associated with the loss of peripheral sympathetic and parasympathetic nervous function²¹.

In seven patients with baroreflex failure, the relative efficacy of phenoxybenzamine (given in a daily dose of 10 to 80 mg orally for two to four days), clonidine (given in a daily dose of 0.3 to 2.4 mg orally for two to four days), and placebo was assessed during hospitalization. The number of blood-pressure surges and the levels of blood pressure and tachycardia during these surges were tabulated.

These studies were performed at the Clinical Research Center of Vanderbilt University. The protocol was approved by the institutional review board of the university, and all study subjects gave informed written consent.

A standard radioenzymatic method was used with catechol-O-methyltransferase for the simultaneous differential assay of norepinephrine, epinephrine, and dopamine.

The results in the different groups were compared by analysis of variance and, when appropriate, unpaired t-tests. The therapeutic responses of the patients were analyzed by paired t-tests. Two-tailed tests were used, and a P value of less than 0.05 was considered to indicate statistical significance. The results are presented as means ±SE.

The defining features of baroreflex failure were an ability of stress (increase in heart rate) or sedation (decrease in heart rate) to modify the heart rate, the absence of a tachycardic response to the hypotensive effect of nitroprusside, and the absence of a bradycardic response to the pressor effect of phenylephrine¹². In terms of these variables there was no overlap between the patients with baroreflex failure and the normal subjects (Table 2).

All patients had labile hypertension, either constantly or episodically. Each of the three patients with sustained hypertension had had two carotid-body tumors (chemodectomas) resected 3 to 20 years apart; none had had hypertension or symptoms of baroreflex failure after the resection of the first tumor. After the resection of the second tumor and the loss of glossopharyngeal- and vagal-nerve function, the patients had sustained hypertension for 24 to 72 hours followed by episodic hypertension.

In the patients in whom baroreflex failure developed more gradually (for example, those in whom it appeared months to years after neck irradiation), there was no initial phase of sustained hypertension, as determined by a review of their history and medical records. Nevertheless, episodic hypertension was prominent and persisted during the six-month to eight-year period of follow-up. Most patients had no decrease in blood pressure on standing (unless they were receiving phenoxybenzamine or were volume-depleted), and in at least four patients systolic blood pressure rose 10 to 30 mm Hg with the assumption of upright posture.

The symptoms of baroreflex failure are shown in Table 1. Headache, tachycardia, diaphoresis, and flushing were generally present only during periods of blood-pressure elevation, which lasted from 3 to 30 minutes, and the highest blood pressures caused the most devastating symptoms. Most patients also appeared to have emotional lability, even between blood-pressure surges.

Figure 1. Figure 1. Blood-Pressure Monitoring in a 43-Year-Old Man Approximately Two Weeks after Surgical Removal of a Carotid-Body Tumor and Five Years after Removal of a Contralateral Carotid-Body Tumor.

While blood pressure was being monitored, the patient's right hand was immersed in ice water for 60 seconds. The blood pressure immediately rose and continued to rise for several minutes after discontinuation of the cold stimulus. Symptoms appeared during this time and then resolved as blood pressure and heart rate returned to normal during the following 30 minutes. On some occasions, the patient had spontaneous paroxysms of similar magnitude.

A typical episode of hypertension induced in a patient by a cold pressor test is shown in Figure 1. The patient's blood pressure was normal during the two hours preceding the test. The blood pressure continued to increase for approximately 10 minutes after the test stimulus was removed, but ultimately returned to base line. Similar hypertensive episodes occurred spontaneously and could be precipitated by even minor mental arousal, such as mental-arithmetic calculations, in all patients, although most were less severe than the one illustrated in Figure 1.

Ten patients reported increased nervousness or emotional lability after the onset of their illness, in response to specific questioning about it. The nervousness was more prominent in the patients who had the greatest elevations in blood pressure, and it increased during blood-pressure elevations. During these elevations, the patients also had a sensation of warmth with pallor (pale flushing), palpitations, and in the most severe episodes, headache and diaphoresis. The constellation of symptoms thus closely resembled that of patients with pheochromocytoma²². The diagnosis of pheochromocytoma was seriously considered at some time in virtually all these patients, but was ruled out by biochemical and radiographic tests and also by the improvement, or at least the absence of an increase, in hypertensive episodes during follow-up.

Figure 2. Figure 2. Variations in Systolic Blood Pressure, Heart Rate, and Plasma Norepinephrine Concentrations in Patients with Baroreflex Failure and Normal Subjects.

The patients and normal subjects were monitored in a metabolic unit at rest and while ambulatory, but no vigorous exercise was permitted. MAX denotes maximum, and MIN minimum. The P values are for the comparison with normal subjects. To convert values for norepinephrine to nanomoles per liter, multiply by 0.005911.

The mean extremes of blood pressure and heart rate during inpatient monitoring are shown in Table 2 and Figure 2. The peak systolic blood pressures were significantly higher in the patients than in the normal subjects, and the patients had lower systolic blood pressures (usually at night) than did the normal subjects. The heart rate generally paralleled blood pressure, with the same wide variations. Two patients had heart rates above 90 beats per minute at all times, possibly reflecting the partial loss of efferent parasympathetic control of the heart rate as a result of damage to the right vagal nerve.

The plasma norepinephrine concentrations changed in parallel with the changes in blood pressure and to a lesser extent with the changes in heart rate (Table 2 and Figure 2). The peak plasma norepinephrine concentrations in the patients were significantly higher than those in the normal subjects. Plasma epinephrine concentrations were also elevated during most attacks, in three patients to more than 200 pg per milliliter (1.1 nmol per liter). Urinary excretion of epinephrine plus norepinephrine averaged 118 μg per 24 hours (697 nmol per 24 hours), more than twice normal (P = 0.015).

The cold pressor test elicited a much greater response in the patients with baroreflex failure than in the normal subjects (Table 2). In some patients, the test initiated a hypertensive paroxysm that continued for many minutes after the removal of the stimulus.

Figure 3. Figure 3. Fall in Systolic Blood Pressure in Response to Clonidine in Patients with Baroreflex Failure and Normal Subjects.

The hypotensive effect of clonidine in the patients was approximately four times that in the normal subjects (Figure 3), but the response to clonidine varied, depending on the initial blood pressure. The clonidine-induced fall in blood pressure was greater during periods of hypertension than during periods of normotension. The plasma norepinephrine concentrations decreased substantially in each of the 10 patients in whom they were measured. The mean (±SE) plasma norepinephrine concentration before clonidine administration was 422 ±141 pg per milliliter (2.49 ±0.83 pmol per liter), and two hours later it had fallen to 180 ±47 pg per milliliter (1.06 ±0.28 pmol per liter, P = 0.021).

The heart-rate response to propranolol was dependent on the prevailing level of sympathetic activation; there was little response when the blood pressure and heart rate were low. During episodes of tachycardia, the heart rate decreased, with a mean fall of 12 beats per minute (range, 8 to 38) in patients with baroreflex failure after they were given propranolol. The heart rate did not increase after the administration of atropine except when patients' blood pressures and heart rates were normal or low. The mean increase in the heart rate in response to atropine was 10 beats per minute (range, 7 to 15).

Table 3. Table 3. Comparison of Blood Pressure, Heart Rate, and Catecholamine Values in Patients with Baroreflex Failure, Autonomic Failure, and Essential Hypertension and in Normal Subjects.

A comparison of the results of biochemical and pharmacologic tests in patients with baroreflex failure, patients with essential hypertension, and patients with autonomic failure caused by the Bradbury-Eggleston syndrome is shown in Table 3. There were significant differences between the patients with baroreflex failure and those with autonomic failure in all variables except the bradycardic response to phenylephrine. There were also substantial differences in the results of most tests in the patients with baroreflex failure and those with essential hypertension. Plasma norepinephrine and epinephrine concentrations were higher and the hypotensive and norepinephrine-lowering effects of clonidine were greater in the patients with baroreflex failure. The most striking difference was in the bradycardic response to phenylephrine, which was the defining characteristic of the group with baroreflex failure.

Table 4. Table 4. Results of Therapy on the Frequency and Severity of Pressor Surges in Seven Patients with Baroreflex Failure.

Some patients with baroreflex failure required constant antihypertensive therapy, so that the efficacy of different drugs could not be systematically assessed. In seven patients, however, it was possible to compare the efficacy of clonidine, phenoxybenzamine, and placebo (Table 4). Clonidine, albeit over a very wide dose range, was effective in reducing both the frequency (P = 0.008) and severity (P = 0.013) of surges of hypertension and tachycardia, whereas phenoxybenzamine attenuated the increase in systolic blood pressure (P = 0.019) without decreasing the heart rate or reducing the frequency of attacks. Three patients had orthostatic hypotension during treatment with phenoxybenzamine, as compared with one patient who was treated with clonidine. Supine blood pressure was lower between as well as during attacks in patients receiving either drug. Eight patients were ultimately treated with clonidine patches. One patient was treated with methyldopa, but an unusually large daily dose (3750 mg orally) was required for blood-pressure control.

Over time, the dose of clonidine could be reduced in many patients. After two to four years, two patients discontinued clonidine therapy and found that their hypertensive episodes could be controlled most of the time with diazepam in a dose of 5 mg three times daily.

The powerful role of baroreflexes in blood-pressure control in humans was shown by studies in which carotid baroreflex function was blocked by infiltrating the carotid-sinus region with procaine^24,25. Bilateral blockade resulted in a substantial increase in systolic blood pressure (approximately 75 mm Hg) and heart rate (approximately 50 beats per minute). During the course of these studies, several subjects had systolic blood pressures exceeding 300 mm Hg, thus providing evidence that when patients with baroreflex nerve lesions were ultimately identified, they would have severe hypertension. Comparable results were obtained in other patients by other investigators^26-28.

We found that clinical baroreflex failure may indeed cause severe, labile hypertension. The correlation of elevations in plasma norepinephrine with pressor episodes suggests that these episodes are caused by unrestrained activation of the sympathetic nervous system. The possibility that vasopressin or other humoral pressor agents might also be released was not addressed in our study. A spectrum of clinical symptoms may accompany the hypertension, including headache, palpitation, a hot sensation, diaphoresis, and emotional lability. Our patients closely resembled the patients described by Kuchel et al.¹⁶ and Aksamit et al.,¹³ whose characteristics are included in Table 2 (Patients 12 and 13, respectively).

There are several reasons for the heterogeneity in the clinical expression of baroreflex failure. Most severe symptoms occur when the interruption in the baroreflex is sudden -- for example, after surgery or injury. Indeed, in the first 24 to 72 hours after such bilateral nerve injury, the hypertension can be constant and may require the continuous infusion of nitroprusside or phentolamine. One patient (Patient 8) had several episodes of apnea during the first 24 hours after surgery, perhaps because oxygen-sensing ability of the carotid body requires the integrity of the glossopharyngeal nerves for its information to be conveyed to the central nervous system. Subsequently, the episodes of hypertension may become less frequent and the associated symptoms less dramatic.

Another explanation for the heterogeneity may be the degree of baroreflex impairment in each patient. No patient had unilateral nerve damage. Even with bilateral nerve damage, there may be differing degrees of involvement of the glossopharyngeal and vagal nerves, which are in close approximation throughout much of their passage through the neck. One previously described patient with baroreflex failure had bilateral functional impairment of the aortic baroreflex fibers carried in the vagal nerves, whereas the cardiopulmonary-reflex fibers, presumably traveling in the same nerve bundles, appeared to have been spared¹³.

Lesser degrees of baroreflex failure have occasionally been detected with hemodynamic monitoring during or soon after carotid endarterectomy or carotid-body surgery^29-31. The true incidence of long-term elevations or lability of blood pressure after such operative procedures is not known.

Three of our 11 patients with baroreflex failure had received radiation therapy to the neck, suggesting that patients undergoing this therapy should be followed for the development of abnormal blood-pressure control. We¹⁷ and Aksamit et al.¹³ have previously described such patients. To our knowledge, no systematic assessment of hypertension in this group of patients has been undertaken.

The two patients whom we classified as having idiopathic baroreflex failure deserve special consideration. Their dysfunction may lie in the brain stem,³² although no abnormality was detected in this structure by computed-tomographic scanning or magnetic resonance imaging in one of these patients (these tests were not done in the second patient). These two patients could have the Page syndrome³³ or a similar disorder^16,34-37. The Page syndrome is characterized by periodic blotchy flushing and perspiration of the face, upper chest, and occasionally the upper abdomen, sometimes in association with cold extremities, headache, tachycardia, and hypertension³³. Vascular compression in the brain stem has also been proposed as a cause of some cases of hypertension³⁸.

It is interesting to speculate why severe orthostatic hypotension does not occur in the face of major damage to arterial baroreflex nerves, despite the suggestion that these reflexes are the principal means by which blood pressure is maintained in the upright posture. The simplest explanation is that some cardiopulmonary-reflex information is still being integrated. Alternatively, some other compensatory mechanism may come into play during the assumption of upright posture. In this regard, cardiovascular-control nuclei in the brain stem might receive information about posture from a number of sensory sources unconnected to the baroreflex. For example, information about position might be obtained from visual cues from the occipital cortex or other sites in the visual pathways, the cerebellum, and the neurovestibular system. These sources collectively might compensate for the absence of baroreflexes.

The clinical presentation of baroreflex failure bears an immediate and striking resemblance to that of pheochromocytoma²². In the 1980s, Bravo et al. proposed that hypertensive patients with high plasma norepinephrine concentrations could be differentiated from those with pheochromocytoma by the ability of clonidine to lower plasma norepinephrine concentrations in the former but not the latter³⁹. Our patients had a 57 percent decrease in plasma norepinephrine after the administration of clonidine, similar to that reported by Bravo et al. in their patients without tumor. These results indicate that this test is useful for ruling out pheochromocytoma. However, as shown in Table 3 and in previous reports,²³ there is some overlap in clonidine responsiveness between patients with baroreflex failure and patients with essential hypertension. This raises the possibility that baroreflex failure may be the cause of labile hypertension in some patients with hyperadrenergic essential hypertension who have high plasma norepinephrine concentrations. This could be readily ascertained by measuring the heart-rate response to phenylephrine.

In conclusion, the manifestations of baroreflex failure in humans range from an acute fulminant hypertensive crisis requiring urgent treatment with nitroprusside to a syndrome of habitual volatility of blood pressure and heart rate with hypertensive surges in response to stress, punctuated by periods of normal or even low blood pressure during rest. Differentiating baroreflex failure from other causes of labile hypertension is essential in devising effective treatment.