8. Medical Differential Diagnosis

In medical patients with severe anxiety, the physician must initially focus on any chronic medical illness that could precipitate anxiety (temporal lobe epilepsy, hypoglycemic episodes in a diabetic, paroxysmal atrial tachycardia in a patient with Wolff-Parkinson-White syndrome) as well as review all pharmacologic agents that the patient is taking that could cause anxiety either as a side effect of normal therapeutic levels or as a marker of toxic blood levels (Rosenbaum 1982). For instance, an asthmatic with a toxic aminophylline serum level and a diabetic with recently increased insulin dosage may both suffer from symptoms that simulate anxiety attacks.

Mackenzie and Popkin (1983) have suggested the term "organic anxiety syndrome" for anxiety symptoms that are found on history, physical examination, or laboratory tests to be secondary to an organic disease, medication, or drugs and where there is no clouding of consciousness (as in delirium), no loss of cognitive function (as in dementia), no predominant disturbance in mood (as in organic affective syndrome), and no hallucinations or delusions (as in organic delusional syndrome or organic hallucinosis).

Medical disorders known to be associated with anxiety are listed in exhibit 9. Among this list of disorders, the medical illnesses that can cause symptoms most resembling panic disorder include hyperthyroidism, cardiac arrhythmias, mitral valve prolapse, hypoglycemic episodes, temporal lobe epilepsy, pheochromocytoma, caffeinism, pulmonary embolus, electrolyte abnormalities, Cushing's Syndrome, and menopausal symptoms. This review focuses on thyroid disease, hypoglycemia, temporal lobe epilepsy, and pheochromocytoma and their differential diagnoses with panic disorder. This chapter also discusses the importance of marijuana, cocaine, amphetamines, and other street drugs in occasionally precipitating the onset of panic disorder and the mimicking of panic disorder by withdrawal from central nervous system depressants and opiates.

Thyroid Disease and Panic Disorder

Klein, in his classic 1964 paper on panic disorder, hypothesized that panic patients fell into two groups: (1) a younger group, age 20 to 40, who develop panic attacks in association with an acute life event, often separation from a significant other, and (2) an older group who develop panic disorder in conjunction with a medical illness, especially an illness with apparent hormonal fluctuations.

Exhibit 9. Medical disorders and substances that can mimic panic symptoms

No definite proof of this division of patients by age of onset and precipitating factors has developed, but there is continued interest in hormonal relationships to panic disorder. Thyroid disease has been mentioned most frequently. Several case report series have described both generalized anxiety and panic disorder developing secondary to hyperthyroidism (Katerndahl and Vande Creek 1986; Turner 1984). Patients with hyperthyroidism often experience anxiety, tachycardia, palpitations, sweating, dyspnea, irritability, diarrhea, and diffuse anxiety (Lesser et al. 1987). Lindemann and colleagues (1984) found, in a retrospective study of 295 panic disorder patients, a high prevalence of thyroid disorders (9.2 percent), based on patient self-reports. Katon and colleagues (unpublished data) also found a significantly higher prevalence of thyroid disease in patients with panic disorder than in controls in a primary care clinic, based on physician reports. However, three recent studies failed to find evidence of abnormally increased thyroid hormone levels in panic disorder patients (Yeragani et al. 1987; Lesser et al. 1987; Stein and Uhde 1988).

One study did show that panic disorder patients had significantly higher levels of free ts and T4 (Adams et al. 1985), although these levels were still in the normal range. Fishman and colleagues (1985) demonstrated a high incidence of very low TSH levels. Taken together, these two studies suggest a relative increase in thyroid function in patients with panic disorder. This relative (but not abnormal) increase would be consistent with the hypothesized increase in adrenergic activity in panic disorder patients, which could increase the synthesis of thyroid hormone (Roy-Byrne et al. 1988). Consistent with this possibility are three studies that found reduced or blunted TSH responses to TRH in panic disorder patients (Roy-Byrne et al. 1986; Hamlin and Pottash 1986; Hamlin 1987) and also in a subgroup of depressed patients (Roy-Byrne et al. 1988).

The above data suggest that the majority of patients with panic disorder are euthyroid, but may have some abnormalities of their hypophyseal-pituitarythyroid axis secondary to increased noradrenergic central activity. These changes are probably not clinically important. Slight elevation of thyroid hormone levels occur in many acute psychiatric illnesses (psychosis, depression, and anxiety), and these elevations represent a stress response rather than thyroid disease (Nusynowitz and Young 1979).

Overall, patients with panic disorder should have a thyroid screen ordered if they have a history of increased appetite, weight loss, heat intolerance, irradiation of neck and chest, subtotal thyroidectomy or iodine 131 treatment, or physical examination findings of ophthalmopathy, goiter, or hypermetabolic state (Raj and Sheehan 1987).

Hypoglycemia

Hypoglycemia can cause symptoms such as cognitive anxiety, tachycardia, sweating, lightheadedness, tremors, and hunger. It may occur from excess insulin in diabetics, secondary to a prediabetic state, or from other medical disorders, such as insulinoma. Primary care physicians frequently see anxious and depressed patients who wonder if hypoglycemia is a cause of their symptoms. This was particularly true in the 1970s, when many lay books and periodicals warned of the symptoms of hypoglycemia, provoking an editorial in the New England Journal of Medicine entitled "Non-hypoglycemia is an epidemic condition" (Yager and Young 1984). The authors described patients who believed they had hypoglycemia but frequently had psychiatric illnesses. The stigmatization of mental illness and the social acceptability of medical illness made the diagnosis of hypoglycemia preferable. This diagnosis enabled patients to deal actively and easily with their symptoms by simply following dietary rituals or prescriptions. Also, a diagnosis of hypoglycemia enabled these patients to attain the sick role, entitling them to increased family and social support.

Hypoglycemia has been carefully studied as a cause of panic disorder, but little supportive evidence has been found. German and Martinez (1984) examined blood glucose in 10 patients with panic disorder during lactate infusion (which precipitates acute anxiety attacks in approximately 70 percent of patients with panic disorder). At the point of panic, the lowest blood sugar reported in any patient was 81 mg/dl. Uhde and colleagues (1984) induced hypoglycemia in nine drug free, panic disorder patients with a standard oral challenge of glucose. Although these patients had quite typical hypoglycemic symptoms of tachycardia, light-headedness, sweating, and free-floating anxiety, they reported their symptoms as different from panic attacks.

The diagnostic criteria for hypoglycemia have been controversial, and the value of the glucose tolerance test (GTT) is limited, because 25 to 48 percent of normal subjects appear to develop hypoglycemia during testing (Raj and Sheehan 1987). Moreover, patients with vague somatic complaints who often have anxiety and/or depression are probably more likely to be referred for a 4-hour glucose tolerance test; a low blood sugar finding is often inappropriately labeled as the cause of their symptoms (Yager and Young 1974; Ford et al. 1976).

The physician should have a high index of suspicion for panic disorder in patients presenting with "hypoglycemia," and the GTT should be reserved for patients with a history of postprandial attacks, panic attacks with accompanying hunger, or gastric surgery or a family or personal history of endocrine adenomas (Raj and Sheehan 1987). The following case demonstrates the patient's misattribution of hypoglycemia as the cause of psychiatric symptoms (panic disorder and major depression).

Pheochromocytoma

Pheochromocytoma is a tumor ofchromaffin tissue, with 80 percent occurring in one adrenal gland, 10 percent bilaterally, and 10 percent outside of the adrenal gland (Landsberg and Young 1987). These tumors are found in 1/1000 autopsies and as the cause of 0.1 percent of all cases of hypertension. Most occur in women between the ages of 20 to 40 years. They are associated with chole-lithiasis in 15 percent of all cases, with multiple endocrine tumors in 10 percent, and with neurofibromatosis in 5 percent (Raj and Sheehan 1987).

The typical clinical features include hypertension (40 percent paroxysmal and 60 percent sustained), palpitations, headache, flushing, and sweating (Raj and Sheehan 1987). Postural hypotension and tachycardia may be noted, and anxiety has been reported as the fourth most common symptom (Manger and Gifford 1982).

Starkman and colleagues (1985) carefully interviewed 17 patients with pheochromocytoma to determine if they experienced anxiety symptoms similar to patients with generalized anxiety disorder or panic disorder. The patients’ subjective descriptions of paroxysmal attacks included headache in 13 (76 percent) patients, sweating and restlessness in 9 (53 percent), and nausea, pounding in the chest, shortness of breath, rapid heartbeat, and an empty, tight feeling in the stomach in 8 (47 percent). These patients generally did not describe any feelings of panic, acute anxiety, or terror. The authors found that although 9 of the 17 patients had some physical symptoms of panic attacks, none met DSM-III criteria for panic disorder. Two had definite generalized anxiety. The authors emphasized that anticipatory anxiety, phobic avoidance, and agoraphobia were not found.

The above data suggest that few patients with pheochromocytoma meet criteria for panic disorder, that physical symptoms usually predominate (headache, sweating, restlessness, nausea, tachycardia), and that phobic behavior and avoidance are rare. Given its rarity, pheochromocytoma workups should be considered only if there is a family or individual history of pheochromocytoma, multiple endocrine adenomas, neurofibromatosis, cholelithiasis, panic attacks with predominant throbbing headache, sweating, or flushing, or physical examination findings of neurofibromata, cafe-au-lait spots, postural tachycardia or hypotension, severe hypertension that is paroxysmal or sustained, or abdominal mass (Raj and Sheehan 1987). Problems in differential diagnosis sometimes occur with somatizing patients who minimize or deny psychologic symptoms and stress, or who present with chest pain or headache and have labile hypertension. Finding phobic behavior and avoidance as well as depressive symptoms may help the physician accurately diagnose panic disorder.

Temporal Lobe Epilepsy and Panic Disorder

Temporal lobe epilepsy (TLE) and panic disorder have several features in common. Both disorders may occur paroxysmally without a precipitory event and may be associated with symptoms of terror, fear, and depersonalization or derealization as well as with autonomic symptoms of diaphoresis, flushing, hyperventilation, and tachycardia (Marshall et al. 1983; Raj and Sheehan 1987). Neuropsychiatric studies have demonstrated that stimulation of specific temporal lobe areas may simulate anxiety. Fear, fright, and feelings of deja vu have been produced by electrical stimulation of the right temporal gyrus. Gloor and colleagues (1982) reported that stimulation of the hippocampus, amygdala, and parahippocampal gyrus produced sensations of fear in humans.

Reports in the neurologic and psychiatric literature describe anxiety attacks associated with underlying mass lesions of the right temporal lobe (Wall et al. 1985; Ghadirian et al. 1986) and occurring after right temporal lobectomy (Wall et al. 1986). A report by Reiman and colleagues (1986) that patients with panic disorder have asymmetry of blood flow in the region of the parahippocampal gyrus on positron emission tomography (PET) scan also suggested that dysfunction of the temporal lobe may be associated with severe anxiety.

Although similarities exist between temporal lobe epilepsy and panic disorder, temporal lobe epilepsy can be differentiated by the loss or altered state of consciousness, the progression into other types of epilepsy (such as grand mal seizures), semipurposeful movements or psychic behavior (deja vuJamais vu), or the presence of hallucinations (Raj and Sheehan 1987). Harper and Roth (1962) found that patients with TLE were more likely to have a history of brain trauma, incontinence, and complete loss of conscious when compared to patients with panic disorder.

Electroencephalograms should not be ordered routinely in patients with panic disorder. They should be reserved for patients with a history of altered consciousness, automatism, head injury, hallucinations, psychic phenomenon with episodes of deja vu or jamais vu, past seizures, or a neurologic deficit on physical examination (Raj and Sheehan 1987). If either history or physical examination is positive, regular and sleep electroencephalograms and neurologic consultation is indicated. If a neurologic deficit is present, a computed tomography head scan and/or magnetic resonance imaging (MRI) should be considered as well.

Illicit Drug Use

In some patients, the first anxiety attack is precipitated by the use of marijuana, cocaine, amphetamines, or hallucinogens (Hillard and Viewig 1983; Aronson and Craig 1986). The frightening cognitive or somatic symptoms that can occur with the use of these drugs may cause the patient to respond with anxiety to this perceived loss of control. Alternatively, some of the physiologic effects of these agents may stimulate specific brain receptors associated with anxiety. Thus, the sympathomimetic effects of cocaine and amphetamines can cause an anxiety response. Cocaine, for example, causes the acute release and blocks the reuptake of serotonin, norepinephrine, and dopamine (Mule 1984); chronic use depletes them. Aronson and Craig (1986) hypothesized that the chronic cocaine-induced depletion of biogenic amines could alter the equilibrium of the noradrenergic system by reducing the activity of inhibitory inputs. This reduced inhibitory input may make the chronic cocaine user susceptible to panic attacks. Marijuana causes a well-documented increase in heart rate, probably by betaadrenergic cardiovascular stimulation (Beaconsfield et al. 1972). This side effect may precipitate severe anxiety in patients.

Withdrawal from sedative-hypnotics, alcohol, or opiates can cause symptoms that are difficult to distinguish from anxiety attacks. In most of these cases, historical data (history of alcohol, barbiturate, benzodiazepine, or opiate use) or physical examination findings ("tracks" from intravenous drug use, tremor, or increased pulse and blood pressure) indicate drug use.

Suggested Medical Workup

How much of a medical workup should the physician complete prior to making a definitive diagnosis of panic disorder? Generally, a thorough biopsychosocial history and physical examination are adequate to make the diagnosis in most patients. This is especially true in the typical case of a relatively young patient (age 18-40) who develops classic anxiety attack symptoms, avoidance behavior, and perhaps depressive symptoms in the context of one or more stressful life events.

Many primary care patients present with somatic symptoms and minimize stressors. In these patients, a brief laboratory screen (complete blood count, blood chemistry panel, and thyroid function tests) in addition to a careful history and physical examination is adequate. This brief screen reassures patients that the clinician has taken their symptoms seriously and has ruled out potential medical problems. It is useful for clinicians to advise these patients prior to the laboratory screen that they believe their symptoms to be secondary to an autonomic nervous system disorder (panic disorder), but that laboratory tests will be useful to exclude medical problems.

At times, a patient's focus on a specific complaint, such as chest pain, dizziness, or epigastric pain or worsening of a medical problem, will necessitate further testing (electrocardiogram, caloric testing, or upper GI series). The clinician must realize the advantages and disadvantages of doing multiple laboratory tests. While tests may accentuate the patient's hypochondriacal tendencies, the additional testing may also decrease patient and physician anxiety about missing a serious illness. Some patients may have both a medical diagnosis and panic disorder, and successful treatment of the medical disorder may still leave a residual panic disorder (Katon and Roy-Byrne 1989).