Paroxysmal Hypertension: The Role of Stress and Psychological Factors

Abstract

This paper reviews the limited literature on paroxysmal hypertension. A case report describes the clinical picture frequently seen in specialty hypertension practice, a patient with paroxysmal or intermittent hypertension who proves not to have a pheochromocytoma. The variety of diagnostic labels given to these patients is reviewed, including pseudopheochromocytoma, panic attacks, and hyperventilation syndrome. The clinical features, pathology, diagnosis, and treatment of these syndromes are outlined. It is proposed that successful management of these patients may be best achieved by collaborative care between a hypertension specialist and a psychiatrist or clinical psychologist with expertise in cognitive‐behavioral panic management, stress‐reduction techniques including controlled breathing, and treating health anxiety. The use of drugs effective for treatment of panic disorder can also be helpful in managing these patients.

A common cause of referral to hypertension specialists is the patient with symptoms and findings of paroxysmal or intermittent hypertension who is studied for a pheochromocytoma with negative results. For every patient referred for a suspected pheochromocytoma, there are many more who have the symptoms but do not have a tumor. Although all practicing hypertension specialists must see large numbers of patients with paroxysmal hypertension, the hypertension literature is largely silent on the issue, except for discussion of pheochromocytoma. It is likely that stress and emotion play a major role in many of these cases, but physicians who treat hypertension are often unfamiliar with the underlying problems and are poorly equipped to deal with them. There is a variety of diagnostic labels given to these patients, which include pseudopheochromocytoma, panic attacks, and hyperventilation syndrome. The purpose of this paper is to review the clinical features, pathology, diagnosis, and treatment of these syndromes.

An example of a patient who fulfilled these criteria and who was recently seen in our practice is described below.

CASE REPORT

The patient was an 85‐year‐old white woman referred for evaluation of paroxysmal hypertension. She had a history of mild hypertension and hyperlipidemia of many years, which had until recently been controlled with a thiazide diuretic and a statin. She was well until January 2007, when she was found to have hyponatremia on a blood test. She was hospitalized, and the diuretic was discontinued and replaced with a calcium channel blocker. At this time, she began to have episodes of paroxysmal hypertension, which she described as starting with epigastric pain, flushing in her face, weakness, light‐headedness, and a feeling of warmth and anxiety. Her blood pressure was typically high during these episodes and could exceed 200 mm Hg systolic. The episodes lasted about 30 to 60 minutes and came on rather suddenly, without any obvious precipitating factors. She was currently having about one episode a day and had also been awoken by them from sleep. Since the initial hospitalization, she had one prolonged hospitalization for 2 weeks, at which time she had an extensive workup for pheochromocytoma and other secondary causes of hypertension. Results of the workups were negative, with the exception of the finding of a stenosis of the right renal artery. A stent was inserted, but the attacks were not affected. Two months later, a second stent was placed in a branch of the left renal artery, again without any noticeable effect on the attacks.

She was treated with clonidine 0.2 mg 3 times a day and carvedilol 6.25 mg twice a day. Subsequently, hydrochlorothiazide 12.5 mg every other day was added. She was advised to monitor her blood pressure at home, and if her pressure was high, she was advised to take extra clonidine or carvedilol. She had taken a small number of home blood pressure readings, which ranged from 113/50 to 200/100 mm Hg.

When we saw this patient, a diagnosis of panic disorder was made. She was treated with a brief session of psychoeducation and cognitive‐behavioral panic management regarding the physiology of panic attacks, the role of hyperventilation as a trigger for physical symptoms, and the role of catastrophic thinking in escalating a vicious cycle of increasing anxiety, increased physical symptoms, and increasing catastrophic interpretations of the symptoms. She was instructed in paced breathing as a stress reduction and symptom management technique and given a CD to guide home practice. She was also prescribed clonazepam 0.5 mg twice daily. Due to the difficulty of her traveling to the clinic, a 2‐month follow‐up appointment was scheduled, and the patient was instructed to call the office if she encountered problems or had further questions. When she was seen for follow‐up, she reported daily practice with the breathing relaxation and occasional use of the clonazepam. After several weeks, she was symptom‐free and had remained so for approximately 1 month. Her blood pressure was generally well controlled, although the home monitoring still showed some spikes.

PSEUDOPHEOCHROMOCYTOMA

The term pseudopheochromocytoma, used to describe paroxysmal hypertension that occurs in the absence of a pheochromocytoma, is of historical interest because it represented the first serious attempt to describe the syndrome. Dr Otto Kuchel, who worked with Jacques Genest in Montreal, popularized the concept in the 1980s. ¹ , ² He described a series of 7480 hypertensive patients referred in Montreal for evaluation, of whom there were 8 who turned out to have a pheochromocytoma, but 688 who had paroxysmal or intermittent hypertension. The paroxysms were accompanied by tachycardia, sweating, flushing, nervousness, shaking, and other manifestations of acute sympathetic activation (Table). He also reported flushing in 74% and panic attacks in 47%, neither of which are characteristic symptoms of pheochromocytoma. Nausea and polyuria were other distinguishing symptoms of this syndrome. In addition, the onset of symptoms was often provoked by external events. In contrast, pheochromocytoma attacks may be quite spontaneous. Measurements of plasma catecholamines showed that plasma levels of free norepinephrine and epinephrine were increased in some patients because of defective inactivation by conjugation, but there was a characteristic increase of plasma dopamine. ² Kuchel suggested that this increase of dopamine explained some of the symptoms that distinguished the condition from pheochromocytoma, such as flushing, anxiety, nausea, and polyuria. Unfortunately, no one seems to have followed up these observations, and measurements of plasma dopamine levels have not been reported in patients with panic disorder.

Although Dr Kuchel's work was featured in a recent book on secondary hypertension, ² this research was performed in the 1980s ¹ and the concept of pseudopheochromocytoma has largely been ignored. The description of panic disorder was first officially introduced in 1980, ³ and there is, in fact, much overlap between what Kuchel described and what was subsequently described as panic attacks. The concept of pseudopheochromocytoma was revived by Mann, ⁴ who described a series of patients with paroxysmal hypertension, in whom the paroxysms were not triggered by stress or panic but who acknowledged a history of severe emotional trauma. In some of the cases, disclosure of this trauma resulted in a cessation of the attacks. Of interest, increased sympathetic arousal and hypervigilance are core symptoms of posttraumatic stress disorder and panic attacks are not uncommon ⁵ ; it is possible that pseudopheochromocytoma is related to these.

PANIC ATTACKS AND HYPERTENSION

Panic disorder has been reported to occur in about 5% of patients seen by primary care physicians, ⁶ and subclinical levels of the symptoms are thought to be considerably more prevalent. ⁷ , ⁸ Although it is commonly considered a benign condition, there is evidence that patients with anxiety disorders are at increased risk for coronary heart disease and sudden death. ⁹ , ¹⁰ . The main diagnostic criterion is recurrent and unexpected anxiety attacks. The typical symptoms occurring during an attack (Table) include palpitations, sweating, trembling, dyspnea, chest or abdominal pain, paresthesia, light‐headedness, and anxiety. ¹¹ , ¹² , ¹³ Studies using ambulatory blood pressure monitoring ¹⁴ , ¹⁵ have clearly documented that blood pressure can increase dramatically within a few minutes, in one case from 110/80 to 220/100 mm Hg. ¹⁵ Although it is often assumed that panic attacks are precipitated by external stress, it is important to note that they most often occur without any obvious precipitating factor and may occur during sleep. ¹⁶

There is substantial evidence that although both hypertension and panic disorder are common in the general population, they occur together more frequently than would be expected by chance. In patients visiting a hypertension clinic, it was found that 12% of patients fulfilled the criteria for a diagnosis of panic disorder. ¹⁷ In another study from the same clinic, ¹⁸ moderate or severe panic attacks occurred in 28% of patients who were compared with hypertensive patients seen in a primary care clinic, in whom the prevalence of panic attacks was 23%, and normotensive primary care patients, in whom it was 12%. These studies did not find that panic disorder was related to the refractoriness of the hypertension. Katon ¹⁹ also reported an increased prevalence of hypertension in patients with panic disorder. Also, in a large cross‐sectional study ²⁰ in which the prevalence of chronic physical diseases were related to coexisting psychiatric disorders (ie, none, affective disorders, substance use, or anxiety), the prevalence of hypertension was the highest (18.2%) in patients with anxiety disorders (which included panic disorder, obsessive compulsive disorders, and phobias), while in the patients with no psychiatric disorder it was 10.2%(P<.01).

A key question is whether panic disorder results from abnormal sympathetic nervous activity, since the signs and symptoms have all the hallmarks of an acute increase. The most direct way of quantifying sympathetic activity is by direct recording of sympathetic nerve firing rates to skeletal muscle. In patients with panic disorder, rates were found to be normal at rest, but there was a marked increase of sympathetic nerve burst amplitude during a panic attack. ²¹ During spontaneous panic attacks, there was an increase of plasma epinephrine but little change in norepinephrine (dopamine was not measured in this study). ²¹ Another study reported increased baseline plasma epinephrine levels in patients with panic disorder (not during an attack), but normal norepinephrine values. ²² Brain norepinephrine turnover is normal in panic disorder. ²³ Production of norepinephrine in the heart is also normal, but there is defective reuptake of norepinephrine back into the sympathetic nerve cells. ²⁴ Another way of examining sympathetic activity is to study the activity of the baroreflex, which regulates both heart rate and vascular resistance. A study of baroreflex sensitivity in patients with panic disorder found that the baroreflex control of heart rate was normal, but the control of muscle sympathetic outflow was enhanced. ²⁵

Esler's group ²³ has suggested that the biochemical basis for panic disorder may be an increased brain serotonin turnover, which was found to be 4 times higher in patients with panic disorder than in normal controls. These findings may explain why selective serotonin reuptake inhibitors (SSRIs) are particularly effective drugs in managing panic disorder.

One of the factors that has limited progress in understanding the specific physiologic mechanisms in panic disorder is the likelihood that there is considerable heterogeneity among patient presentations ¹¹ and probable causal mechanisms. For example, Klein ²⁶ and Leor and colleagues ²⁷ proposed at least 2 subtypes of panic disorder, that with particular sensitivity to respiratory dysregulation and that which appears to be more cognitively mediated, a model that has received considerable recent validation. ²⁸ , ²⁹ Patients also vary in the degree to which lactate infusion (which induces acidosis) or carbon dioxide inhalation (which induces alkalosis) trigger panic attacks. ³⁰

THE ROLE OF HYPERVENTILATION

Although hypertension is officially regarded as being an asymptomatic condition, anyone who spends time treating hypertensive patients knows that this is not always the case. Patients who have symptoms from whatever cause are more likely to seek medical care and to show up in hypertension clinics as well as in any other medical setting. The symptoms are often vague and difficult to relate to any specific cause. Some years ago, Dr Norman Kaplan wrote an article entitled, “Anxiety‐Induced Hyperventilation: A Common Cause of Symptoms in Patients With Hypertension,” ³¹ in which he described a series of 300 patients referred to him because their hypertension was difficult to treat. The constellation of symptoms that many of these patients described included paresthesia, light‐headedness, dizziness, palpitations, and headache (Table). Of these patients, one‐third had anxiety‐induced hyperventilation diagnosed. The case for attributing these symptoms to hyperventilation was made by establishing that voluntary hyperventilation replicated the symptoms and that rebreathing into a paper bag made them disappear.

The role of hyperventilation in actually causing the symptoms is not entirely clear. A number of studies have shown signs of respiratory dysregulation in patients with panic disorder and hyperventilation syndrome, ³² , ³³ , ³⁴ with particularly consistent findings regarding slow partial pressure of carbon dioxide (pCO₂) recovery after hyperventilation ³⁵ and irregular tidal volume of breathing. ³⁴ On the other hand, a double‐blind study in which patients hyperventilated either on room air or a carbon dioxide‐enriched mixture that prevented any change in end‐tidal pCO₂ found that patients were generally unable to distinguish between the two. ³⁶ This is an area in which extensive individual differences have been found (ie, habituation to the symptoms caused by pCO₂ inhalation). ³⁷ At any rate, it is clear that a simple hyperventilation challenge is unlikely to lead to a reliable diagnosis in many patients.

The blood pressure response to hyperventilation is variable. A classic early study of the hemodynamic effects of voluntary hyperventilation concluded that hyperventilating for 1 minute lowered peripheral resistance by 45% and mean blood pressure by 23 mm Hg. ³⁸ However, in a study of patients with normal blood pressure who were asked to hyperventilate for 5 minutes, blood pressure increased by 9/8 mm Hg and heart rate by 36 beats/min. ³⁹ Fontana and associates ⁴⁰ classified both normotensive and hypertensive individuals into 3 groups according to the blood pressure response to hyperventilation: those who showed an increase, a decrease, or no change. The same group also performed ambulatory blood pressure monitoring in healthy patients and hypertensive patients who were classified according to their response to hyperventilation. ⁴⁰ Although there were no significant differences in the daytime average ambulatory blood pressure between the 3 hyperventilation groups in either the normotensive or hypertensive patients, the peak ambulatory blood pressure levels and the variability of blood pressure during the day were higher in both the normotensives and hypertensives in the group that showed a pressor response to hyperventilation; there was a tendency for the variability of blood pressure to be higher. Martinez and coworkers ⁴¹ found that in patients with panic disorder who developed panic symptoms during hyperventilation, there was an increase in both systolic and diastolic pressure, whereas in healthy patients and those with a diagnosis of panic disorder who did not have symptoms there was a decrease. Taken together, these findings suggest that blood pressure typically increases during hyperventilation in patients with panic disorder and that this may increase the variability of blood pressure during the day.

CAN THESE CONDITIONS BE LINKED?

The symptoms of the various types of paroxysmal hypertension, including pheochromocytoma, are remarkably similar, as shown in the Table. Thus, headaches, palpitations, sweating, anxiety, and nausea, among others, are reported to occur in all of them. Weight loss is also common in pheochromocytoma, but not in the other conditions; this may be attributed to the sustained as opposed to intermittent increase of catecholamine production. Orthostatic hypotension is also common in patients with a pheochromocytoma, but not in the other conditions. On the other hand, flushing is not described as being characteristic of a pheochromocytoma, whereas it is common with the others. About 60% of patients with a catecholaminesecreting tumor appear pale during an attack. ⁴²

An issue that has received little attention is whether all the “nonpheochromocytoma” syndromes are different facets of the same stone. We live in a stressful world, and situations that provoke anxiety are all too common. Individuals who are anxious are more likely to seek medical care; this would explain the much higher prevalence of panic disorder in medical clinics than in the general population. ¹⁷ , ¹⁸ However, the fact that it is highest in a hypertension clinic raises the possibility that there is a causal connection between panic disorder and hypertension. One important issue here is whether the apparent association between panic disorder and hypertension can be explained by increased anxiety associated with an office visit and the production of an exaggerated white‐coat effect. However, the only study to examine the white‐coat effect in patients with panic disorder found that it was not consistently higher than in other patients. ⁴³

The role of hyperventilation is not clear. Hyperventilation can be both a cause of and a response to anxiety. One measure of tendency to hyperventilate is a low resting end‐tidal pCO₂ level. Studies of patients with panic disorder show that while they tend to have lower end‐tidal pCO₂ levels than healthy controls, similar levels have been reported in patients with generalized anxiety disorder. ⁴⁴ Others, however, have found that end‐tidal pCO₂ levels are lowest in patients with panic disorder. ³² , ⁴⁵ Thus, the hyperventilating patients may well have had generalized anxiety disorder or panic disorder.

The occurrence of hyperventilation is only part of the story, however, because the direction of the blood pressure response may also be critical. The results of Fontana and associates ⁴⁰ suggest that individuals who show the most marked pressor response to hyperventilation may be the ones who have the highest peaks of blood pressure during the day and hence are the most likely to have paroxysmal hypertension diagnosed.

EVALUATION OF THE PATIENT WITH PAROXYSMAL HYPERTENSION

All patients with paroxysmal hypertension should be screened for pheochromocytoma, which can be done with the measurement of urine ⁴⁶ or plasma metanephrines. ⁴⁷ If results of these are negative and the symptoms are consistent with the clusters shown in the Table, evaluation for panic disorder and hyperventilation is appropriate. An evaluation and follow‐up with a psychiatrist and/or clinical psychologist with whom the medical practice has a relationship can be extremely helpful. He or she should have experience in somatic presentations of anxiety disorders and experience with cognitive‐behavioral panic management treatment and stress reduction techniques. The importance of emphasizing collaborative care in these patients is essential to getting them to follow through on such a referral. A brief explanation of the physiologic interactions can help reassure the patient that he or she is not being dismissed.

One of the problems is that there has been no specific definition of paroxysmal hypertension, as opposed to normal blood pressure variability. One of the consequences of the wider use of home blood pressure monitoring is the realization of the large spontaneous variability of blood pressure. This is particularly true of elderly patients, in whom spontaneous short‐term blood pressure variability increases and heart rate variability decreases. ⁴⁸ In a study of healthy Japanese patients of different ages, it was found that the standard deviation of the 24‐hour systolic pressure might be as low as 10 mm Hg in a 30‐year‐old, but ≥25 mm Hg in a 60‐year‐old. ⁴⁹ In our experience, differences in consecutive self‐monitored blood pressure readings of 60 to 70 mm Hg in older hypertensive patients are not at all uncommon, so it is hard to distinguish spontaneous variability from the effects of a panic attack. Thus, the definition of paroxysmal or intermittent hypertension becomes problematic. At the present time, there are no established guidelines for the normal range of spontaneous blood pressure variability. Another problem is that in susceptible patients the finding of one high reading may increase anxiety and lead to a maintenance or exacerbation of the blood pressure elevation.

As pointed out by Kaplan, ³¹ numerous visits to the emergency department and extensive workups for secondary causes of hypertension may further augment anxiety in these patients. In the patient we have described above, it is worth noting that her paroxysmal symptoms did not begin until after her first hospitalization and change in therapy.

TREATMENT

Treatment should be focused on 2 areas: controlling the pressure and treating the underlying panic disorder or anxiety state. While the role of antihypertensive drugs is of course critical, as in any other type of nonsurgical hypertension, it should be remembered that the primary effect of drugs is to lower the overall level of blood pressure without affecting its short‐term variations. In general, antihypertensive drugs lower the average level of blood pressure without having much effect on the short‐term variations. There is some evidence to suggest that sympathetic blocking drugs may reduce short‐term blood pressure variability,⁵⁰ and it has been suggested that combined α‐ and β‐blockade may be beneficial in such patients,⁵¹ although this is not well documented. The chief objective of drug treatment should be to maintain the average level of blood pressure while achieving a balance between excessive peaks and troughs. Home monitoring is helpful in this respect, providing that the patient performs readings on a fixed time schedule, rather than only during paroxysms.

Treatment of panic disorder includes both the use of medications (principally clonazepam and SSRI s) and behavioral counseling. Despite the apparent heterogeneity of panic disorder, the drugs and cognitive‐behavioral treatment packages for panic management have been effective for many of these patients presenting in the medical setting. A number of studies have tested a combination of the following components: 1) psychoeducation regarding the psychophysiology of hyperventilation and panic in contributing to symptom induction; 2) stress reduction techniques, such as paced breathing; 3) addressing the catastrophic cognitions and health anxiety that have resulted from the alarming symptoms and their interactions with the health care system during episodes; and 4) interoceptive exposure (in which symptoms are deliberately induced by various physical maneuvers, including intentional acute hyperventilation, to desensitize the patient to the fear associated with the symptoms and reduce avoidance behavior.⁵² We have found in the hypertension practice setting that helping patients identify and work with the respiratory dysregulation patterns has been particularly important. The work includes breathing instruction in session that directs the patient toward breathing that is more focused in the diaphragm than in the upper chest, paced slowly (<10 breaths/min), and limited in tidal volume. The patients' tendency when calling attention to breathing is frequently to take a “deep breath” that is too large and results in overbreathing. The panic management treatment approaches have reduced symptom episodes significantly, with results that persist substantially up to 2 years.⁵²

CONCLUSIONS

Paroxysmal hypertension that is not attributable to a curable secondary origin such as pheochromocytoma is not uncommon, and it is possibly becoming more frequently identified with the increased use of self‐monitoring. It is unclear at what point the normal variability of blood pressure becomes paroxysmal, but it is a source of anxiety to patients and, in some cases, may become a vicious cycle. An evaluation of stress, anxiety, and panic attacks in such patients may be very rewarding and enable treatment to be directed not only at lowering the blood pressure but also at relieving the associated psychological problems.