Pheochromocytoma: an update on genetics and management
- Division of Endocrinology, Second Propedeutic Department of Internal Medicine, Medical School, Hippokration Hospital, Aristotle University of Thessaloniki, Thessaloniki 54642, Greece 1Department of Clinical Biochemistry,, Royal Free University College Medical School, Royal Free Hospital, University College London (University of London), Pond Street, London NW3 2QG, UK
- (Correspondence should be addressed to D P Mikhailidis; Email: mikhailidis{at}aol.com)
Abstract
Pheochromocytomas (PHEOs) are rare neoplasms that produce catecholamines and usually arise from the adrenal medulla and are considered to be an adrenal paraganglioma (PGL). Closely related tumors of extraadrenal sympathetic and parasympathetic paraganglia are classified as extraadrenal PGLs. Most PHEOs are sporadic, but a significant percentage (∼25%) may be found in patients with germline mutations of genes predisposing to the development of von Hippel–Lindau disease, neurofibromatosis 1, multiple endocrine neoplasia type 1 (MEN1) and 2 (MEN2), and the PGL/PHEOs syndrome, based on the described mutations of the genes for succinate dehydrogenase subunit D (SDHD), B (SDHB), and C (SDHC). As one out of four PHEOs turns out to be a hereditary clinical entity, screening for genetic alterations is important, as it provides useful information for a rational diagnostic approach and management. This review discusses the genetics, the pathophysiology of hypertension, the clinical picture, the biochemical and imaging diagnosis, and the preferred therapeutic approach for PGLs/PHEOs. Furthermore, it emphasizes the need for genetic testing in cases with apparently sporadic PHEOs.
Introduction
The adrenal medulla and ganglia of the sympathetic nervous system (SNS) are derived from the embryonic neural crest. The endocrine cells of this sympathoadrenal system synthesize and secrete catecholamines and exhibit a characteristic histochemical (chromaffin) reaction when treated with oxidizing agents. Pheochromocytomas (PHEOs), according to the World Health Organization (2004) classification, are rare tumors arising from catecholamine-producing cells in the adrenal medulla – an intraadrenal paraganglioma (PGL). Adrenal and extraadrenal PGLs produce significant amounts of catecholamines and give rise to the well-known clinical picture of PHEOs. In contrast, the parasympathetic PGLs (mainly in head and neck) rarely produce significant amounts of catecholamines.
During the last few years, a considerable amount of new data, concerning the genetics of PHEO/PGL, have accumulated and changed the whole approach to such patients. It has been shown that in about 25% of cases, PHEO/PGLs develop secondary to germline mutations in any of five susceptibility genes (Dluhy 2002, Neumann et al. 2002, Amar et al. 2005a, Gimenez-Roqueplo et al. 2006, Pacak et al. 2007).
A classic PHEO, a solitary tumor of the adrenal medulla, reminds us of a ‘tip of an iceberg’ because this expression suggests that beyond a single tumor there is potentially a broader clinical picture awaiting exploration. This review summarizes the important relevant data related to this fascinating clinical entity.
Prevalence
The prevalence of PHEOs is not precisely known. Among the general population in Olmsted County, Minnesota, USA, a PHEO occurs in about 1 or 2 per 100 000 adults per year (Beard et al. 1983). This figure suggests that if 20% of the adult population is hypertensive, approximately five PHEOs would be expected among 100 000 hypertensives each year. In a large series of patients screened biochemically for suspicion of PHEO, the incidence has been reported to be as high as 1.9%, occurring equally in men and women (Smythe et al. 1992). On the other hand, a large number of patients with PHEO have nonclassic symptoms such as abdominal pain, vomiting, dyspnea, heart failure, hypotension, or sudden death, suggesting that the majority of PHEOs are not diagnosed during life (Sibal et al. 2006).
Etiology–genetics
Most PHEOs are sporadic and arise from chromaffin cells of the adrenal medulla, but in about 9–23% of cases tumors develop from extraadrenal chromaffin tissue and are then referred as PGLs (Bravo & Tagle 2003, Sibal et al. 2006). In children, multifocal and extraadrenal PHEOs are found in 30–43% of cases (Khafagi et al. 1991, Levine & DiGeorge 2000). Malignant PHEOs account for up to 26–35% of cases (Whalen et al. 1992, Mundschenk & Lehnert 1998) and the prevalence of malignancy in sporadic adrenal PHEOs is 9% (Bravo & Tagle 2003). About 10% of patients with PHEOs present with metastatic disease at the time of their initial work-up (Goldstein et al. 1999). Approximately up to 25% of patients with apparent sporadic PHEOs, without family history of the disease, may in fact be carriers of germline mutations predisposing to the development of neurofibromatosis 1 (NF1), von Hippel–Lindau (VHL), multiple endocrine neoplasia type 1 (MEN1) and 2 (MEN2), and the PHEO/PGL syndromes based on the described mutations of the genes for succinate dehydrogenase subunit D (SDHD), B (SDHB), and C (SDHC) (Baysal et al. 2000, Neumann et al. 2002, Amar et al. 2005a).
Neurofibromatosis type 1 (NF1)
PHEO constitutes a minor but established feature of NF1. This is an autosomal-dominant disorder, with a PHEO frequency estimated at 0.1–5.7% but rising to 20–50% in hypertensive patients with NF1 (Riccardi 1981, Zoller et al. 1997, Bausch et al. 2007). The clinical diagnosis of NF1 requires two of the following seven criteria (Gutmann et al. 1997): six or more café-au-lait spots; two or more cutaneous neurofibromas or a plexiform neurofibroma, inguinal, or axillary freckles; two or more benign iris hamartomas (Lish nodules); at least one optic-nerve glioma; dysplasia of sphenoid bone or pseudoarthrosis; and a first degree relative with NF1, according to the preceding criteria.
NF1 is caused by inactivating mutations of neurofibromin, a tumor suppressor gene, which encodes a GTPase-activating protein involved in the inhibition of Ras activity, which controls cellular growth and differentiation. The susceptibility gene, the NF1 gene, is localized to chromosome subband 17q11.2 (Gutmann et al. 1994). In most cases, PHEOs are benign (90%) and single (84%), followed by bilateral (10%) and sympathetic PGLs (6%; Walther et al. 1999a, Bausch et al. 2006). Most of them occur in adults and produce predominantly norepinephrine (NE) and therefore present with hypertension and noradrenergic symptomatology.
VHL disease
VHL is an autosomal-dominant disease with an incidence of 1 in 3600 births (Maher et al. 1991). PHEO develops in ∼20% of patients with VHL, with a mean age at onset in the second decade of life, although such tumors often occur even later. VHL is caused by mutations of the VHL gene localized to chromosome 3p25–26. A germline mutation in the VHL gene predisposes carriers to tumors in multiple organs (Lonser et al. 2003). The gene is a tumor suppressor that encodes a protein (pVHL) which is involved in blood vessel formation by regulating the activity of hypoxia inducible factor (HIF)-1α (Iliopoulos et al. 1996). This protein inhibits the accumulation of hypoxia-induced proteins through ubiquitin-mediated degradation of HIF-1α, under conditions of normoxemia (Hes et al. 2003). In carriers of VHL gene mutations, the regulation of genes such as the vascular endothelial growth factor and other genes involved in cellular growth seems to be lost, predisposing the VHL carriers to both benign and malignant tumors in multiple organs. Loss of pVHL function reduces HIF degradation and increases vascular endothelial growth factor that leads to angiogenesis (Kaelin 2002). These tumors may include hemangioblastoma in the retina (also referred to as retinal angioma); cerebellum and spine; renal cell carcinoma (clear cell type); PHEO; islet tumors of the pancreas; endolymphatic sac tumors; and cysts and cystadenoma in the kidney, pancreas, epididymis, and broad ligament (Choyke et al. 1995). If present, metastases from renal cell carcinoma and neurological complications from cerebellar hemangioblastomas are the most common causes of death.
On the basis of its clinical expression, VHL disease has been divided into four subtypes with a central role for PHEO (Maher et al. 1996, Koch et al. 2002, Hes et al. 2003). Patients with VHL type 1 disease have loss of pVHL function, due to gene deletions or specific missense mutations and develop retinal or central nervous system hemangioblastomas and renal carcinoma, but they are not at risk for PHEO. Patients with type 2 disease have mainly specific missense mutations and develop hemangioblastoma and PHEO. They are also at low (type 2A) or high (type 2B) risk for renal cell carcinoma. A small percentage of patients with VHL (type 2C) will have only PHEO without the other tumors. In this type, the missense mutation has the gain of function effect causing PHEO, but these particular missense mutations do not substantially affect pVHL-mediated HIF degradation. Thus, even the loss of the second VHL gene would not lead to development of hemangioblastoma or renal cell carcinoma in patient with a type 2C mutation. Knowledge of the VHL mutation could tailor clinical attention and surveillance to the organs at risk and potentially reduce the psychological anxiety and the cost of unnecessary investigations.
PHEO may present as the first or only manifestation of VHL (VHL type 2C; Hes et al. 2003). Consequently, VHL carriers can present as apparently sporadic PHEO. VHL catecholamine-producing tumors are most commonly PHEOs (90%), although sympathetic PGLs have been described. Approximately half of PHEOs are bilateral and most produce NE (Eisenhofer et al. 1999, 2001b). There are uncommon examples of malignant catecholamine-producing tumors in VHL, frequently sympathetic PGLs (Pujol et al. 1995).
MEN1
MEN1 is an autosomal-dominant syndrome characterized by primary hyperparathyroidism, pancreatic islet cell neoplasms, and pituitary adenomas caused by inactivating mutations of the MEN1 locus coding for the suppressor protein menin localized to chromosome 11q13. MEN1 is very rarely associated with PHEO, all unilateral, rarely malignant, and most characterized by hypertension and predominant NE production (Schussheim et al. 2001).
MEN2
MEN2 is an autosomal-dominant syndrome caused by activating mutations in the RET proto-oncogene located on chromosome 10q11.2, which encodes a transmembrane receptor tyrosine kinase involved in the regulation of cell proliferation and apoptosis, causing constitutive activation of the receptor. As a result of tissue-specific expression, calcitonin-producing parafollicular cells and adrenomedullary chromaffin cells initially undergo hyperplasia with a high rate of subsequent neoplastic transformation. MEN2A is characterized by medullary thyroid carcinoma (MTC), hyperparathyroidism, and PHEO. MEN2B is characterized by MTC, mucosal ganglioneuromas, and PHEOs (Gagel et al. 1988, Brandi et al. 2001). PHEO occurs in approximately half of gene carriers and is almost always located within the adrenal glands.
Bilateral PHEOs occur in approximately half of patients with MEN2 who have PHEO and are usually confined to the adrenal medulla; their development is frequently asynchronous, with separation by as much as 15 years (Recasens et al. 2007). PHEOs occur most commonly with codon 634 (MEN2A) or 918 (MEN2B) RET proto-oncogene mutations (Eng et al. 1996). Malignant PHEOs are uncommon, <5%, and are generally found in patients with large tumors (Chevinsky et al. 1990). The pattern of catecholamine production in MEN2 PHEO differs from that seen in other hereditary forms of PHEO. Epinephrine (E) is produced in disproportionately large amounts, resulting in an early clinical phenotype characterized by attacks of palpitations, nervousness, anxiety, and headaches rather than the more common patterns of hypertension seen with sporadic or other hereditary tumors (Hamilton et al. 1978, Gagel et al. 1988, Eisenhofer et al. 1999, Yip et al. 2003, Young & Abboud 2006).
Paragangliomas (PGLs)
PGLs are rare tumors that arise from chromaffin cells. They represent 10–18% of all chromaffin tissue-related tumors that are reported at a rate of 2–8 cases/million a year. According to their origin, they are classified as sympathetic or parasympathetic (Lack 1997).
Sympathetic PGLs
They are derived from the sympathetic chain and usually are located in the chest, abdomen, or pelvis. The clinical picture is a consequence of either the secretion of catecholamines or the size of the tumors with consequent impingement on other structures. The morphology of adrenal PHEOs and extraadrenal PGLs is usually very similar. The frequency of malignancy is much higher in sympathetic tumors with extraadrenal location.
Parasympathetic PGLs
They are tumors of the parasympathetic ganglia usually found in the head and neck region, arising from the cell nests located adjacent to blood vessels, such as the carotid body or the ganglion jugulare. Unlike PGLs in the abdomen, they are usually biochemically silent, and malignancy is seen in <10% of the cases (Lack 1997).
In a large series of 236 patients with 297 benign PGLs evaluated at the Mayo Clinic during 1978–1998, the mean age at diagnosis was 47±16 years (Erickson et al. 2001). Of the 297 PGLs, 205 were in the head and neck region and 92 were below the neck. They were discovered and diagnosed incidentally on imaging in 9% of patients. The most frequent PGLs in the neck were carotid body tumors, and those most common below the neck were abdominal periaortic–pericaval tumors. The most frequent symptoms for the patients with head and neck tumors were palpable neck mass (55%) and cranial nerve palsies (16%). The clinical presentations of these tumors were dominated by local mass effects (neck mass, tinnitus, and cranial nerve dysfunction) and only a small proportion (4%) was hyperfunctional. In patients with PGLs below the neck, one or more of the classic catecholamine excess pentad of headaches (26%), palpitations (21%), perspiration (25%), pallor (12%), and orthostasis (6%) were observed; hypertension was present in 64%. Their location were periaortic and pericaval, perirenal, mediastinal, intracardiac, pulmonary parencymal, intraspinal, sacral, duodenal, jejunal, pancreatic, or in the organ of Zuckerland, bladder, or prostate (Neumann et al. 2004, Benn et al. 2006). Retroperitoneal PGLs are most likely to be malignant and present with pain or a mass. They tend to metastasize to the lungs, lymph nodes, and bones or they can extend locally and may destroy adjacent vertebrae and can cause spinal compression (O'Riordain et al. 1996, Edström Elder et al. 2003). Disease-causing mutations in three genes (SDHB, SDHD, and SDHC) encoding-subunits of succinate dehydrogenase of mitochondrial complex II are responsible for most cases of familial PGLs. SDH has an important function in the Krebs cycle and the mitochondrial respiratory chain.
PGL–SDHD
PGL–SDHD is an autosomal-dominant syndrome characterized by familial and isolated head and neck parasympathetic PGLs and less frequently by sympathetic PGLs and PHEOs, caused by mutations in SDHD gene located on chromosome 11q21–23 (Baysal et al. 1997, Baysal et al. 2002, Benn et al. 2006). Head and neck PGLs are generally regarded as benign tumors but can extend into the skull. In addition, SDHD mutation carriers should be observed for multifocal tumors, infrequent malignancy, and the possibility of extraadrenal PHEOs (Benn et al. 2006). The hereditary nature of the disease may be masked by maternal imprinting of SDHD, resulting in only paternal transmission of SDHD-associated disease (Baysal et al. 2002), although recently a different mechanism has been proposed to account for exclusive paternal transmission (Hensen et al. 2004). PHEOs may be unilateral or bilateral and the mean age of diagnosis is 43 years.
PGL–SDHB
PGL–SDHB is an autosomal-dominant syndrome characterized by sympathetic extraadrenal PGLs and malignant disease (Brouwers et al. 2006, Timmers et al. 2007a,b). The syndrome is caused by inactivating mutations in the tumor suppressor SDHB gene located on chromosome 1p35–36 (Astuti et al. 2001). These mutations cause mitochondrial complex II destabilization and may activate the hypoxic/angiogenic pathway predisposing to tumor formation, with a very strong association with a malignant intra- or extraadrenal phenotype. Malignant PHEOs are reported in 35% in these patients (Neumann et al. 2004). In a series of 44 patients with malignant PGLs, SDHB mutations were found in 30% and nearly one-third had metastases originating from an adrenal primary tumor and two-thirds from an extraadrenal tumor. The latter patients had a frequency of SDHB mutations of 48%. This high frequency of SDHB mutations justifies the priority for SDHB germline mutations testing in these patients (Brouwers et al. 2006).
In another recent study of 29 patients with SDHB-related abdominal or thoracic PGLs, the mean age of diagnosis was 33.7±15.7 years, 76% had hypertension and 90% lacked a family history of PGL. All primary tumors, but one, originated from extraadrenal locations. Symptoms were related to tumor mass effect rather to catecholamine excess and the predominant biochemical phenotype consisted of hypersecretion of NE/dopamine, with 10% of tumors being silent (Timmers et al. 2007b). Therefore, SDHB mutation in a patient with single initial tumor is a prognostic factor for malignancy (Havekes et al. 2007). In families with SDHB mutations, maternal imprinting has not been noted. There is also an increased risk for renal cell carcinoma and papillary thyroid cancer (Schiavi et al. 2005).
Other PGLs
A PGL caused by mutations in SDHC gene located on chromosome 1q21 has been described as an autosomal-dominant syndrome, without maternal imprinting. It is characterized by benign and seldom multifocal head and neck parasympathetic PGL (Schiavi et al. 2005). Another autosomal-dominant syndrome with familial head and neck parasympathetic PGLs has been described exclusively in children of fathers carrying the gene which was mapped to chromosome 11q13.1 but not identified yet (Mariman et al. 1995, Baysal et al. 1997).
Genetic testing
Hereditary catecholamine-producing PHEOs and PGLs can be caused by germline mutations in any of the five above-described genes and mutation testing is now routinely available for the four of the above genes. Extensive clinical experience demonstrates that germline mutations are responsible for about 25% of cases instead of 10% thought to be hereditary previously (Dluhy 2002). Most importantly, 7.5–27% of tumors without an obvious syndrome or family history result from otherwise unsuspected germline mutations in one of these four genes. Despite these high figures of unsuspected germline mutations, most experts agree that testing must be restricted to patients who fulfill several clinical criteria (Amar et al. 2005a, Gimenez-Roqueplo et al. 2006, Jimenez et al. 2006 Pacak et al. 2007).
Family history
Familial PHEOs/PGLs syndromes are inherited in an autosomal-dominant manner; thus, an affected individual has a one in two (50%) chance of passing on the mutation to each child. A family history may be found in patients with VHL-, MEN2-, NF1-, and SDHD-related syndromes. In families with SDHD mutation, inheritance is complicated by the maternal imprinting phenomenon. It must be reminded that carriers of the SDHB or SDHD gene mutations do not necessarily have evidence of tumors. Those mutations have an age-related penetrance, and the lifetime risk of developing PGLs approaches 100% by the age of 70 years (Benn et al. 2006). In the absence of any family history, sudden deaths should be recorded, bearing in mind that 50% of catecholamine-producing tumors remain undiagnosed until death.
Age in presentation
Hereditary tumors usually occur at a younger age than sporadic tumors, but the age range is wide being 5–69 years for the mutation carriers and 4–81 years in the group of patients with no identified tumors (Gimenez-Roqueplo et al. 2006). Genetic testing is more necessary in young adults, especially for VHL disease, as 36% of PHEOs/PGLs in children occur secondary to germline mutations (Barontini et al. 2006).
Tumor location – metastases
Patients with SDHB/SDHD gene mutations commonly present with extraadrenal often multifocal disease. SDHB gene mutations carry a high risk of malignant disease and therefore testing for such mutations is warranted and identification of a mutation in the SDHB gene is a high risk factor for malignancy. In contrast, malignant disease and extraadrenal tumors are rare in MEN2, so testing for RET gene mutations is not rewarding; furthermore, it is inappropriate to test for RET gene mutations in tumors characterized by an increase in urinary or plasma NE but not E. This differs from PHEOs in VHL syndrome which do not produce significant amounts of E (Table 1).
Familial pheochromocytomas
Pathophysiology of hypertension
PHEO, even if they do not secrete, synthesize catecholamines at increased rates that may be up to 27 times the synthetic rate of the normal adrenal medulla. This persistent overproduction is due to lack of feedback inhibition on tyrosine hydroxylase. Catecholamines are then produced in quantities that greatly exceed the vesicular storage capacity and accumulate in the cytoplasm. Catecholamines in the cytoplasm are subject to intracellular metabolism; the excess catecholamines and their metabolites diffuse out of the PHEO cells into the circulation (Fitzerald & Goldfren 2004). In contrast to the normal adrenal medulla, PHEO cells ordinarily contain more NE than E. The PHEO of the adrenal medulla in adults contain both catecholamines, whereas the extraadrenal PGLs rarely secrete E.
Traditionally, the hypertension of PHEO has been thought to result solely from the action of circulating catecholamines on cardiovascular circulating receptors. The activity of the SNS was thought to be normal or depressed reflexly because of baroreceptor resetting, and neurally released NE was considered of minor physiological importance in comparison with the effects of markedly elevated plasma catecholamine levels. However, clinical and experimental studies suggest that the SNS is intact, markedly enhanced and that SNS function is integral to the maintenance of elevated blood pressure (BP) in this form of catecholamine-induced hypertension (Langer et al. 1980, Johnson et al. 1983). The elevation of sympathetic activity during elevation of circulating catecholamines is postulated to be due to loading of sympathetic vesicles with catecholamines, increased sympathetic neuronal impulse frequency, and selective desensitization of presynaptic α2-adrenergic receptors, which results in enhanced release of neuronal NE during nerve stimulation. Therefore, because of enhanced SNS activity and excessive stores of NE in sympathetic nerve terminals, any direct or reflexly mediated stimulus to the SNS could release excessive quantities of NE into the synaptic cleft and produce a hypertensive crisis. The easier access of NE released from the postganglionic neuron at its receptor site on effector cells can result in marked symptoms with relatively small increments in circulating NE. These findings account for the observation that spontaneous or evoked crises in BP can occur without additional increases in the elevated plasma catecholamine levels and that BP may be normal despite high circulating catecholamine levels. This may account for the paroxysms of hypertension that are triggered by pain, emotional upset, intubation, anesthesia, or surgical skin incision and explain the elevations in serum and urine catecholamines that can occur for 10 days or longer after a successful surgical resection of PHEO (Bravo et al. 1990, Bravo & Tagle 2003).
Many PHEOs secrete significant amounts of neuropeptide Y (NPY) along with NE. NPY has potent direct and indirect cardiovascular effects. It increases coronary and peripheral vascular resistance independently of α-adrenergic mechanisms by interacting with vascular G-protein-coupled receptors (O'Hare & Schwartz 1989). In some vascular beds, NPY has no direct vasoconstrictor effects but potentiates NE-induced vasoconstriction (Macho et al. 1989). NPY appears to contribute to hypertension in most patients with PHEO. In contrast, few PGLs secrete NPY (deS Senanayake et al. 1995). PHEOs secrete many other peptides, many of which contribute to clinical symptoms, such as PTHrP (hypercalcemia), adrenocorticotrophin (ACTH; Cushing's syndrome), erythropoietin (erythrocytosis), and IL-6 (fever). Most PHEOs secrete chromogranin A which can be assayed as a tumor marker, and many secrete other peptides which have not been documented to produce clinical manifestations (Grossrubatscher et al. 2006).
Clinical picture
The clinical picture is characterized by diverse manifestations reflecting the variations of hormone secretion, the patterns of release, and the individual-to-individual differences in catecholamine sensitivities (Eisenhofer et al. 2004, Amar et al. 2005b, Reisch et al. 2006, Sibal et al. 2006). As mentioned before, there is no correlation between circulating levels of catecholamines and hypertension. Plasma catecholamines were increased in periods of normotension in some patients and sudden increases in BP may not be encountered with further elevations in plasma catecholamines. In general, the hypertension is paroxysmal in 48% of patients, persisted in 29%, and 13% have normal BP (Bravo et al. 1979). NE-secreting tumors are usually associated with sustained hypertension. Tumors that secrete large amounts of E together with NE are associated with episodic hypertension. Pure E-producing tumors can produce hypotension rather than hypertension (Fitzgerald & Goldfren 2004). Large (∼50 g) cystic PHEOs are often asymptomatic because the secreted catecholamines are metabolized within the tumor, and only a small amount of free catecholamines is released into circulation. Symptoms typically include a sudden rise of BP with concurrent episodes of headache (80%), diaphoresis (70%), and palpitations (60%). The episodes usually last minutes or hours; symptoms usually begin abruptly and subside shortly. These episodic paroxysms may not recur for months or may recur many times daily. Each patient tends to have a different pattern of symptoms, with the frequency of severity of episodes usually increasing over time. In one study, this symptomatic triad was found to have a sensitivity of 90.9% and specificity of 93.8% (Plouin et al. 1981). Other symptoms may include anxiety (50%), a sense of dread, tremor, or paresthesias. However, about 8% of patients may be completely asymptomatic, usually those with familial forms of the disease or with large, cystic tumors (Kudva et al. 1999). Approximately 5% of adrenal incidentalomas have proved to be PHEOs and diagnosis relies on imaging phenotype and measurement of fractionated metanephrines and catecholamines in 24 h urine (Young 2007). The ‘typical’ clinical signs and symptoms occur more frequently in patients with benign tumors. Abdominal pain and dorsalgia occur more frequently in malignant PHEOs and short history and extraadrenal localization are suspicious for malignancy (Glodny et al. 2001). High preoperative 24-h urinary dopamine levels, high tumor weight, elevated tumor dopamine concentration, and postoperative persistent arterial hypertension are factors that increase the likelihood of malignant PHEO (Houbert et al. 1999).
Attacks can occur spontaneously or may occur with bladder catheterization, anesthesia, and surgery. Attacks can be induced by seemingly benign activities, such as bending, rolling over in bed, exertion, abdominal palpation, or micturition (with bladder PGLs). Other disorders can dominate the clinical picture such as hypercalcemia, Cushing's syndrome, thyroid carcinoma, diabetes mellitus, or acute abdomen (Yip et al. 2003). Cardiovascular episodes, such as shock, myocarditis, dilated cardiomyopathy, cardiac arrhythmias, pulmonary edema, and heart failure, and neurological disorders, such as altered mental status, stroke, seizures, focal neurological signs, and symptoms, may also dominate the clinical picture and be the main cause of death (Sutton et al. 1981, Sibal et al. 2006). The administration of nonselective β-blocker therapy without preceding α-blockade in a patient with PHEO may precipitate a crisis with hemodynamic collapse (Sloand & Thompson 1984). The opposite phenomenon has been described in an interesting case report. The administration of unopposed α-blockade induced a state of β-adrenergic overstimulation with tachycardia, diastolic dysfunction, diffuse edema, and hypotension with peripheral vasodilation. The patient responded to small doses of 5 mg propranolol (Kantorovich & Pacak 2005).
A percentage of patients with PHEOs may present with minor or no signs and symptoms and some patients remain undiagnosed despite advances in diagnostic techniques. Elderly patients appear to present a special diagnostic challenge. A contributory factor to the delay in the antemortem diagnosis of PHEOs in the elderly may be a decrease in baroreceptor function with age, or concomitant diseases, the signs and symptoms of which can confound the diagnosis.
Patients with sustained hypertension usually exhibit episodes of orthostatic hypotension and even syncope, which are due to vasomotor receptor desensitization or diminished intravascular volume. Patients with tumors secreting E may present with sinus tachycardia and peripheral vasoconstriction; the radial pulse can become thready or even nonpaplable during a hypertensive crisis. Vasoconstriction is also responsible for the pallor and mottled cyanosis that can occur with paroxysms of hypertension. Reflex vasodilatation usually follows an attack of hypertension and can cause facial flushing. After an intense and prolonged attack of hypertension, shock may ultimately occur. This may be due to loss of vascular tone, low plasma volume, arrhythmias, or cardiac damage. In malignant PHEOs, metastases are usually functional and can cause recurrent hypertension and symptoms, many months or years after surgery that was thought to be curative. Metastases can cause a variety of problems as space-occupying lesions (Bouloux 2002, Bravo 2004). PHEOs or secreting PGLs may recur after initial surgery. The risk of recurrence is fairly high and these patients should be followed up indefinitely. The risk is higher in younger patients, with larger tumors and was more likely to have familial disease or bilateral extraadrenal or right-sided tumors (Amar et al. 2005b). On the other hand, no recurrence was found in 71 patients with benign tumors followed up for 144 months (Edström Elder et al. 2003).
Biochemical diagnosis
The optimal approach to biochemical confirmation of catecholamine-secreting tumors is debatable. When performed under appropriate clinical settings, currently available tests can establish the diagnosis in more than 95% of cases. Given that PHEOs are a heterogeneous group of hormone-secreting tumors with variable metabolism, it is prudent to recommend that for a high diagnostic accuracy, multiple tests should be performed. The combination of different tests may increase sensitivity and specificity. Readily available biochemical tests include 24-h urinary catecholamines (NE and E) and vanillyl-mandelic acid (VMA), 24-h urinary total and fractionated metanephrines (normetanephrine and metanephrine), plasma concentrations of NE and E, and plasma concentrations of free metanephrines.
Whereas levels of free catecholamines are increased by even minimal anxiety and stress, levels of metanephrines are much less affected. Therefore, plasma catecholamines and metanephrines must be measured in a supine patient at rest for at least 20 min after insertion of an indwelling cannula in the forearm (Lenders et al. 2002, Eisenhofer et al. 2003) or in a seated ambulatory patient by standard venipuncture (Sawka et al. 2003). However, measurement of plasma metanephrines in blood samples collected from patients in the supine position preserves the high diagnostic sensitivity of the test (Lenders et al. 2007).
Urinary free catecholamines and metabolites
To detect the presence of a PHEO, 24 h urinary free NE and E and their metabolites (normetanephrine, metanephrine, and VMA) are commonly assayed. The demonstration of urinary NE >170 μg/24 h, of urinary E more than 35 μg/24 h, urinary total metanephrines at least 1.8 mg/24 h, and urinary VMA at least 11.0 mg/24 h makes the diagnosis highly probable (Bravo & Tagle 2003, Kudva et al. 2003). Large cystic tumors may release mainly metabolized catecholamines into the circulation as reflected by a relatively high ratio of metabolites to free catecholamines in urine (Crout & Sjoerdsma 1964).
Plasma catecholamines
The majority of patients with hormonally active PHEOs have elevated plasma levels of both NE and E, many exclusively NE and a much smaller proportion exclusively E. These differences in plasma catecholamine levels reflect differences in the expression of phenylethanolamine N-methyltransferase (PNMT), the enzyme that converts NE to E (Feldman et al. 1979). In a large series of 93 patients with either adrenal (n=80) or extraadrenal (n=13) PHEO, only three patients had plasma NE values that fell within the 95% upper confidence limits for values obtained in 104 patients with essential hypertension (i.e., 811 pg/ml). None had values that fell within the 95% upper confidence limits for gender- and age-matched 58 normotensive subjects (i.e., 402 pg/ml; Lenders et al. 1993). However, 30% of patients with adrenal and 62% of patients with extraadrenal PHEOs had plasma E values below the 95% upper confidence limits obtained in 104 patients with essential hypertension (i.e., 135 pg/ml; Lenders et al. 1993).
BP must be recorded during plasma sampling for catecholamine measurements. PHEO cannot be excluded if normal plasma catecholamine values are obtained when the patient is normotensive and asymptomatic. However, normal plasma catecholamine levels in a hypertensive and symptomatic patient make the diagnosis of a PHEO highly unlikely.
Plasma metanephrines
Newer techniques, such as liquid chromatography, allow the separate measurement of normetanephrine, the O-methylated metabolite of NE, and metanephrine, the O-methylated metabolite of E. Most methods also allow additional measurements of methoxytyramine, the O-methylated metabolite of dopamine (Eisenhofer et al. 2005). Measurements of the fractionated metabolites, although they are not widely available, are superior to measurements of total metanephrines in that they allow better detection of tumors that produce predominantly or only one of the three O-methylated metabolites (Eisenhofer 2003, Rosano et al. 1991, Gardet et al. 2001, Václavík et al. 2007).
PHEOs contain considerable amounts of the membrane-bound form of catechol-O-methyltransferase (COMT), resulting in local metabolism of catecholamines to free metanephrines (Eisenhofer et al. 1998). It has been shown that over 90% of circulating metanephrine and between 23 and 40% of circulating normetanephrine are produced by the metabolism of the parent catecholamines within the adrenal glands (Eisenhofer et al. 1995a,b). This makes the adrenals the single largest tissue source of circulating metanephrines in the body, with production surpassing by far that of metanephrines by the liver (Eisenhofer et al. 1995a).
In a large series of 208 patients with PHEOs, 23% had normal plasma NE concentrations compared with only 4% with normal levels of normetanephrine (Eisenhofer et al. 2003). Similarly, 68% of patients had normal plasma concentrations of E compared with only 47% with normal levels of metanephrine. Plasma levels of normetanephrine <112 pg/ml and of metanephrine <61 pg/ml virtually exclude PHEO. With plasma concentrations of normetanephrine above 400 pg/ml or of metanephrine above 236 pg/ml, the probability of PHEO is so high that the immediate task is to locate the tumor. When intermediate values are obtained, further investigation with pharmacologic tests is needed.
In another study, only 1 of 33 patients with PHEO had normal levels of both normetanephrine and metanephrine, a patient with a dopamine-secreting PGL (Sawka et al. 2003). Such tumors are extremely rare and are usually found as extraadrenal PGLs. Predominance of dopamine and relative lack of production of the other catecholamines in such tumors are due to deficiency in tumor cells of dopamine-β-hydroxylase, the enzyme that converts dopamine to NE (Eisenhofer et al. 2005).
A multicenter cohort study examined the sensitivity and specificity of several biochemical tests in 214 patients with proven PHEOs (138 sporadic, 48 VHL, 23 MEN2, 3 NF1) and in 644 patients in whom the diagnosis was suspected but excluded (Lenders et al. 2002). In this study, the sensitivity of plasma free metanephrines was clearly superior to plasma catecholamines (99 vs 92% in sporadic PHEOs, and 97 vs 69% in hereditary PHEOs). The specificity of free metanephrines was also better in hereditary PHEOs (96 vs 89%), but was relatively low (82%), although superior to plasma catecholamines (72%). Test sensitivity for urinary fractionated metanephrines was similar to plasma free metanephrines, but lacked specificity for both hereditary (82%) and sporadic (45%) disease. In sporadic PHEOs, urinary catecholamines had excellent sensitivity (97%) and total metanephrines the best specificity (89%). The combination of different tests did not improve the diagnostic yield beyond that of a single test of plasma free metanephrines. The advantage of plasma metanephrines is that they are continuously produced and released by the tumor, in contrast with plasma catecholamines which are released intermittently. The above results have led the authors to recommend against the use of multiple biochemical tests to exclude PHEO in favor of a single test of plasma free metanephrines (Fig. 1).
Algorithm for the biochemical and imaging diagnosis of pheo/PGL. Pheo, pheochromocytoma; PGL, paraganglioma; CT, computed tomography; MRI, magnetic resonance imaging; MIBG, metaiodobenzylguanidine; PET, positron emission tomography.
The combination of resting plasma catecholamines (NE plus E) at least 2000 pg/ml and urinary metanephrines (normetanephrine plus metanephrine) at least 1.8 mg/24 h has a diagnostic accuracy close to 98% in both sporadic and hereditary PHEOs (Bravo & Tagle 2003). Assays for plasma catecholamines and urinary metanephrines have the lowest false-negative rates (7%), and assays of urinary NE and E the next higher (14%). Urinary VMA measurements have high false-negative rates (41%) and should not be used for screening purposes (Bravo & Tagle 2003). However, all four tests have excellent specificity when elevated. When available, the measurement of plasma free metanephrines should be performed, especially when hereditary PHEO is suspected. However, whether the measurement of plasma free metanephrines should be the sole diagnostic test for PHEO remains to be determined.
Serum chromogranin-A (CgA)
CgA has been suggested as an alternative diagnostic test to catecholamines for the detection of PHEO, because neither its secretion nor its measurement is influenced by drugs commonly used for the treatment of hypertension in PHEO patients. However, mild degrees of renal impairment can lead to significant increases in serum concentration of CgA, because the kidneys play a major role in its clearance. Therefore, although CgA is relatively sensitive (86%) in the diagnosis of PHEO, it has poor diagnostic specificity. When combined with elevated plasma catecholamines in patients with creatinine clearance of at least 80 ml/min, its diagnostic specificity increases to 98% (Canale & Bravo 1994, d'Herbomez et al. 2007).
Clinical situations and medication-associated false-positive results
When testing for PHEO, some consideration should be given to possible causes of false-positive results, including clinical situations, medications, inappropriate sampling conditions, and diet (Table 2). Biochemical testing for PHEO should ideally be carried out after discontinuation of medications known to influence catecholamine levels and their metabolites or interfere directly with biochemical analyses. Tricyclic antidepressants are the established cause of false-positive results, probably due to the inhibition of NE reuptake (Esler et al. 1991, Veith et al. 1994, Young 1997). Phenoxybenzamine (POB), a nonspecific α-adrenoreceptor blocker commonly used in the treatment of patients with PHEO, may lead in high rates of false-positive results (Eisenhofer et al. 2003). False-positive elevations of plasma metanephrines and catecholamines with selective α1-adrenoreceptor blockers, such as doxazosin, are not a problem (Eisenhofer et al. 2003), while with calcium channel blockers appear restricted to NE, an effect most likely due to the reflexive sympathetic activation occurring with the short-acting agents (Wenzel et al. 1997).
Disorders that may increase both plasma and urinary catecholamines metabolites to levels often seen in pheochromocytoma
Drugs that inhibit central sympathetic outflow, such as clonidine, methyldopa, and bromocriptine, decrease plasma catecholamines in normal and hypertensive subjects, but have little effect on the excessive catecholamine secretion by PHEO. Labetalol can increase plasma catecholamines and urinary metanephrines determined by HPLC (high pressure liquid chromatography) to values seen in PHEO (Feldman 1987). Acetaminophen can cause spurious increases in plasma free metanephrines. In addition, a urinary metabolite of buspirone, an anxiolytic drug, is artificially measured as metanephrine, resulting in marked increase in measured metanephrine excretion (Cook et al. 1995).
Pharmacologic tests
Basal levels of plasma catecholamines are usually several-fold higher in patients with PHEO than in other subjects, even taking into account normal variations due to postural change, exercise, and emotional arousals. Basal total plasma catecholamine levels ≥2000 pg/ml is diagnostic of PHEO, while <500 pg/ml essentially rules it out. Concentrations in between, especially those exceeding 1000 pg/ml in medically stable patients, suggest the need for further testing and confirmation by pharmacologic tests, either provocative or suppressive.
Provocative tests
A provocative test, with administration of histamine or glucagon, is rarely employed when the clinical findings are highly suggestive of PHEO, but the BP is normal or slightly elevated and plasma catecholamines are between 500 and 1000 pg/ml (Bravo & Gifford 1984). A positive glucagon stimulation test requires at least a 3.0-fold increase and/or more than 2000 pg/ml in total plasma catecholamines. This test has high specificity (100%) but low sensitivity (81%). In addition, the glucagon stimulation test has proven insensitive in familial forms of PHEO (Walther et al. 1999b). Close monitoring of BP is required and phentolamine must be at hand for the treatment of severe hypertensive crises.
Suppression tests
Suppression tests seem more physiologic and safer than provocative tests. The most widely used is the clonidine suppression test, which was introduced by Bravo et al. (1981) to address the problem of how to distinguish patients with PHEO from those with false-positive biochemical results after initial testing for the tumor. Clonidine, by activating α2-adrenoreceptors in the brain and the sympathetic nerve endings, suppresses NE release from sympathetic nerves. Therefore, decreases in elevated plasma NE levels after clonidine suggest sympathetic activation, whereas a lack of decrease suggests PHEO.
The clonidine suppression test is carried out after an overnight fast with the patient supine. Plasma catecholamines and metanephrines are measured before and 3 h after a single oral dose of clonidine (0.3 mg/70 kg, adjusted for body weight as necessary; Eisenhofer et al. 2003). A normal response is defined as a fall of plasma catecholamines from baseline of at least 50% and below 500 pg/ml. When the test is performed in patients with plasma catecholamines of at least 1000 pg/ml, the false-positive and false-negative rates are 2% (Bravo & Tagle 2003).
Another strategy to improve the sensitivity and specificity of the clonidine suppression test is to measure plasma normetanephrine. In hypertensive patients without PHEO, normetanephrine consistently decreases after the administration of clonidine. On the other hand, plasma normetanephrine levels do not fall and remain elevated after clonidine in 96% of patients with PHEO compared with only 67% for NE. Therefore, this metabolite may be a better endpoint marker for the clonidine suppression test than the parent amine, because PHEO cause larger, more consistent, and less episodic increases of normetanephrine than of NE (Eisenhofer et al. 1998, 2001a, Raber et al. 2000).
A remaining minor limitation is that responses of plasma metanephrine cannot be used to distinguish true- from false-positive results for this metabolite. This is due to the fact that over 90% of circulating metanephrine is normally derived from metabolism of E within adrenal chromaffin cells, a process that is independent of E release (Eisenhofer et al. 1995a,b). Indeed, plasma metanephrine concentrations do not decrease after clonidine in patients without PHEO (Gutmann et al. 1994). In addition, clonidine-induced changes of E offer limited help (Gross et al. 1987, Mannelli et al. 1987).
Localization
Only after the diagnosis of PHEO has been confirmed by biochemical studies, localization of the tumor should be attempted combining at least two imaging techniques (Fig. 1). Anatomical imaging studies, such as computed tomography (CT) and magnetic resonance imaging (MRI), should be combined with functional imaging studies, such as nuclear medicine methods, in order to locate adrenal, extraadrenal, recurrent, or metastatic PHEOs (Ilias & Pacak 2004). The clinicopathological behavior of these tumors may help localize them more precisely. Adrenal tumors are common in patients 60 years or older, are rarely associated with extraadrenal tumors, and may be bilateral when occurring in patients with familial syndromes. Therefore, age and the presence of family history are important considerations when determining the type and location of PHEO. The majority of tumors (95%) occur within the abdomen (Bravo 1994). The most common extraadrenal sites are the superior and inferior paraaortic areas (75%), the bladder (10%), the thorax (10%), and the head, neck, and pelvis (5%) (Whalen et al. 1992).
Anatomical imaging
CT and MRI are common initial imaging modalities providing high sensitivity but less than optimal specificity (Ilias & Pacak 2004, Qiao et al. 2007). They should be carried out over the abdomen first, because PHEO are mostly situated within the adrenal medulla. If no adrenal masses are seen, attention should be focused initially to the paraspinous area. The majority of PGLs occurs in the paraaortic region or around the renal hilum and may be visible on CT/MRI (Mayo-Smith et al. 2001).
Adrenal PHEOs of 0.5–1.0 cm or larger or extraadrenal PHEOs at least 1.0–2.0 cm in size can be detected by CT preferably with scanning sections of 2–5 mm thickness (Quint et al. 1987, Pacak et al. 2001c,d). CT densitometry may help to differentiate adrenal adenomas from metastases (Dunnick & Korobkin 2002). Small PHEOs (1–2 cm in diameter) are usually homogeneous in appearance, with soft tissue density (∼40–50 Hounsfield units) and show uniform enhancement with contrast. Larger PHEOs may undergo hemorrhage and can be unhomogeneous, while areas of low density can be seen after tumor necrosis (Dacie & White 1993, Sohaib et al. 2001). If the CT of the abdomen and pelvis is negative, a CT of the chest and neck should be performed (Braedel et al. 1986, Bender et al. 1997). Spiral CT is preferred for the detection of small thoracic tumors (Shin et al. 1986). If the PHEO tumor has been localized, there is no need to proceed to MRI, but functional imaging is required to confirm that the tumor is indeed a PHEO and to rule out metastatic disease.
If the CT is negative, in a patient with biochemically proven PHEO, MRI should be performed. MRI, with or without gadolinium enhancement, provides the highest sensitivity among current imaging techniques. On MRI T1 sequences, PHEOs have a signal like those of liver, kidney, and muscle and can be differentiated easily from adipose tissue. PHEOs appear hyperintense to the liver or muscle on T2-weighted image, whereas benign tumors appear isointense. However, such intense signals can be elicited by hemorrhages or hematomas, adenomas, and carcinomas, so an overlap with PHEOs must be considered and specific additional imaging is needed to confirm that the tumor is PHEO (Varghese et al. 1997, Mayo-Smith et al. 2001, Prager et al. 2002). MRI should be substituted for CT in children, pregnant women, and situations where radiation exposure must be minimized (Witteles et al. 2000).
Functional imaging
Adrenal masses are present in about 5–9% of the general population (Angeli & Terzolo 2002, Grumbach et al. 2003, Thompson & Young 2003). In patients without known primary cancers, about 90% of incidental masses are benign. In patients without clinical evidence of adrenal hyperfunction, about 85% are nonfunctional. Nonetheless, every adrenal mass should be evaluated to rule out hypersecretion, malignancy, or metastasis to the adrenal gland. The prevalence of malignancy in sporadic adrenal PHEOs is 9% (Bravo & Tagle 2003) and about 10% of patients with PHEOs present with metastatic disease at the time of their initial work-up (Goldstein et al. 1999). Given that there is no consensus on the existence of absolute clinical, laboratory, or imaging criteria to predict malignancy and multiplicity of PHEOs (van der Harst et al. 2000, Thompson 2002), the need to exclude metastatic disease or multiple tumors is important. This need might be fulfilled with functional imaging modalities using various radiopharmaceuticals that provide physicians with whole body, PHEO-specific, scans.
PHEO cells usually abundantly express specific catecholamine plasma membrane and vesicular transporter systems, enabling imaging with metaiodobenzylguanidine ([131I]MIBG) and [123I]MIBG scintigraphy, as well as several positron emission tomography (PET) ligands (Fig. 1). The functional imaging test of choice is [123I]MIBG scintigraphy (sensitivity, 83–100%; specificity, 95–100%), or, if this is not available, then [131I]MIBG scintigraphy should be performed (sensitivity, 77–90%; specificity, 95–100%; Bravo 1994, Nielsen et al. 1996, Furuta et al. 1999, van der Harst et al. 2001, Cecchin et al. 2006, Pacak et al. 2007).
If the MIBG scan is negative, PET imaging using 6-[18F]-fluorodopamine (DA), [18F]-dihydroxyphenylalanine (DOPA), [11C]-hydroxyephedrine, or [11C]-epinephrine are new promising, specific radionuclide techniques for localization of PHEOs (Shulkin et al. 1992, Pacak et al. 2001a, Ilias & Pacak 2004, Kaji et al. 2007). [18F]DA has been proven to be an excellent agent to localize adrenal and extraadrenal PHEOs, including metastatic lesions (Pacak et al. 2001a). [18F]DA is also more specific for the diagnosis of PHEOs than other amines such as DOPA, because the latter are taken up as amines by all body cells and are converted to DA (Ilias & Pacak 2004). In cases of aggressive PHEOs, which can be both [123I]MIBG scintigraphy and 6-[18F]-fluorodopamine negative, [18F]-fluorodeoxyglucose PET ([18F]-FDG) may be useful in the diagnostic localization of metastatic PHEO (Shulkin et al. 1999, Mamede et al. 2006, Timmers et al. 2007a). FDG PET is not recommended for initial diagnostic localization, because its sensitivity is slightly less than that of MIBG and its specificity is considerably lower due to uptake of FDG by a variety of other neoplastic and nonneoplastic processes (Shulkin et al. 1999).
If the PET study is also negative, the patient probably has an unusual type of PHEO (in which tumor cells do not express the NE transporter system or may have a low number of catecholamine storage granules), or malignant PHEO, and scintigraphy nonspecific ligands, such as somatostatin receptor, scintigraphy with octreoscan should be carried out (Fig. 1). Somatostatin receptor scintigraphy, with either [123I]Tyr3-octreotide or [111In]DTPA-octreotide, has been used in patients with PHEOs (Krenning et al. 1993, Lastoria et al. 1995, Lauriero et al. 1995, Tenenbaum et al. 1995, Kopf et al. 1997, Limouris et al. 1997). Octreoscan studies may be negative in most patients with benign PHEOs, despite positive [123I]MIBG or [131I]MIBG studies (Maurea et al. 1996, van der Harst et al. 2001). Malignant or metastatic PHEOs are better detected with octreoscan compared with [123I]MIBG (87 vs 57%) (van der Harst et al. 2001), because MIBG as well as [18F]DA are sometimes negative in patients with malignant PHEOs, possibly due to decreased expression of the cell membrane NE transporter by less well-differentiated cells (Ramachandran et al. 1993, Houben et al. 1994).
Venous sampling coupled with measurement of catecholamines or, preferably, free metanephrines to localize the tumor through the discovery of a secretory gradient is an ultimate modality to be used with caution in selected cases, where all imaging have failed (Newbould et al. 1991, Chew et al. 1994, Pacak et al. 2001d). If this is technically impossible, a repeat noninvasive localization work-up after 2–6 months is a more attractive and preferable choice.
Management
After precise localization, surgical removal of the tumor should follow with the expectation that all symptoms will be relieved in the majority of patients with benign PHEOs and in the hope that metastatic spread will be limited in the minority with malignant ones. Appropriate antihypertensive drugs are used to manage hypertension, prevent cardiovascular complications, and diminish the magnitude of postoperative hypotension.
The preoperative use of the nonselective α-blocker POB was mainly advocated to counteract the sudden release of catecholamines during surgical intervention (Ross et al. 1967, Perry & Gould 1972). The dose is increased over a period of 14 days, starting with 10 mg b.d. Under tight BP control, the dose is increased by 10 mg per day up to 1 mg/kg per day, in three divided doses (Kinney et al. 2002). However, POB produces significant orthostatic hypotension and reflex tachycardia, while it may prolong and contribute to the hypotension that follows operation. Yet, despite adequate α-blockade, total elimination of cardiovascular disturbances is seldom achieved, and hypertensive crises have been reported in patients during manipulation of the tumor (Desmonts & Marty 1984). It has also been reported that similar perioperative complications occurred whether or not patients received preoperative α-blockers (Newell et al. 1988, Boutros et al. 1990). In another study, fewer perioperative complications were observed in those not given α-blockers (Ulchaker et al. 1999).
Selective postsynaptic α1-receptor antagonists, such as prazosin, terazosin, and doxazosin, have been used to circumvent some of the side effects of POB. Since these drugs leave the presynaptic receptors on the neuronal surface open, they do not produce reflex tachycardia (Prys-Roberts 2000). In addition, they have a shorter duration of action, permitting more rapid adjustment of dosage and a reduced duration of postoperative hypotension (Bravo & Tagle 2003). Gradual increase of dose from 1 to 16 mg once a day is necessary.
β-Blockers should be given to control tachycardia and arrhythmias, but only after α-blockers have been started. Nonselective β-blockers lead to the loss of β2-receptor-mediated vasodilatation and the unopposed effects of α-receptors cause vasoconstriction leading to arterial hypertension and increased afterload, causing myocardial infarction and pulmonary edema (Wark & Larkins 1978, Sloand & Thompson 1984, Baysal et al. 2000).
Calcium channel blockers have been used alone or with selective α1-receptor blockers to successfully control BP and symptoms in patients with PHEO (Serfas et al. 1983, Lenders et al. 1985, Takahashi et al. 1989). They relax arteriolar smooth muscle and decrease peripheral vascular resistance by inhibiting NE-mediated release of intracellular calcium and/or transmembrane calcium influx in vascular smooth muscle (Lehmann et al. 1983). These drugs do not produce hypotension and therefore may be used safely in patients who are normotensive but have occasional episodes of paroxysmal hypertension. In addition, they are useful in managing cardiovascular complications because they prevent catecholamine-induced coronary vasospasm and myocarditis. Hypertensive crises are controlled with nitroprusside, nitroglycerine, or phentolamine.
The combination of CT or MRI with MIBG scintigraphy has contributed to the precise localization of PHEO. In addition, the availability of new drugs and procedures to safeguard intraoperative hemodynamics and the introduction of innovative surgical techniques have dramatically changed the surgical approach to PHEO.
Since its first description in 1992, laparoscopic adrenalectomy has become the preferred method for the removal of the tumor (Gagner et al. 1992). However, for the time being, there are few studies comparing intraoperative hemodynamic stability in laparoscopic versus conventional open surgery with heterogeneous results (Fernandez-Cruz et al. 1996, Inabnet et al. 2000, Sprung et al. 2000, Edwin et al. 2001, Kazaryan et al. 2004, Kim et al. 2004). There are no randomized trials. In two studies, the intraoperative hemodynamic values during laparoscopic adrenalectomy were comparable to those of traditional open surgery (Sprung et al. 2000, Weismann et al. 2006). In the past, the upper limit for endoscopic resection of PHEO was 6 cm; however, as experience has grown, many authors currently report excisions up to 11 cm (Kalady et al. 2004). The advantages of the minimal invasive techniques are shorter hospital stay and recovery time, decreased requirements for analgesics, a cosmetically better result, and patient satisfaction. In cases of bilateral sporadic or hereditary PHEO, adrenal sparing surgery preserving adrenocortical function should be favored (Neumann et al. 1999).
Patients with PHEO have a high plasma volume requirement during and after surgery. Expansion of intravascular volume either with plasma volume expanders or 2 l normal saline the evening before surgery and the generous replacement of blood loss during the procedure greatly reduce the frequency and the severity of postoperative hypotension. It is well known that intraoperative hypertensive episodes are related to a sudden and marked catecholamine release from the tumor (Joris et al. 1999). In most cases, this is a consequence of direct tumor manipulation (Fernandez-Cruz et al. 1996), but it may also be triggered by the pneumoperitoneum of the laparoscopy approach due to its mechanical effect on the tumor (Tauzin-Fin et al. 2004).
The principal responsible factor for the postoperative severe hypotension seems to be the shrunken blood volume, which is no longer supported by intense vasoconstriction. In addition, the sudden decrease of catecholamines in the immediate postoperative period leads to an increase in insulin secretion and simultaneous decrease in the formation of glucose from glycogen and fat, resulting in hypoglycemia. Biochemical evidence that the tumor is successfully removed cannot be obtained immediately, because plasma catecholamines remain high for some days (Bravo et al. 1979). This is probably due to the emptying of extratumoral pools of catecholamines and should not be interpreted only as a partial response to surgery. The normalization of plasma and urinary catecholamine and metanephrine concentrations should be checked 10 days after surgery. If these concentrations remain high, [123I]MIBG scintigraphy should be performed and may disclose distant metastases, for which MIBG uptake was masked before surgery by the higher metabolic activity of the primary tumor (Plouin & Gimenez-Roqueplo 2006). PHEO is considered a curable cause of hypertension. If the BP remains high after operation, some of the tumor may have been left behind, or a renal artery may have been injured, with induction of renovascular hypertension.
Since PHEO can recur, patients must be followed up for the rest of their lives, particularly if they have inherited or extraadrenal tumors (Lenders et al. 2005). The prognosis is usually excellent for benign PHEOs. In a large series, the risk of recurrence was 3.4-fold higher with a familial disease, 3.1-fold higher with a right rather than a left tumor, and 11.2-fold higher with extraadrenal tumors (Amar et al. 2005b). In patients with one of the familial syndromes, repeated catecholamine assays in combination with plasma calcitonin levels and palpation of the neck for medullary thyroid cancer should be continued for life.
Management of malignant PHEO
Malignant PHEO is compatible with prolonged survival, with symptom-free intervals varying from months to decades. Metastases may appear many years after removal of an apparently benign tumor, and there is at present no certain way to predict which tumors will progress to malignancy. Prognostic factors that have been suggested for future malignancy include large tumor size, local tumor extension at the time of surgery, and the DNA ploidy pattern with DNA diploid being benign and DNA aneuploidy and tetraploidy having a more aggressive nature (Nativ et al. 1992). It has also been reported that the expression of inhibin/activin βB-subunit was absent or weak in malignant PHEO, while it was strong or moderate in almost all benign adrenal tumors (Salmenkivi et al. 2001). Another study suggested that the lack of NPY mRNA expression may differentiate malignant from benign PHEOs (Helman et al. 1989). Various other molecular markers have been proposed, such as expression of human telomerase reverse transcriptase, heat shock protein 90, seretogranin II-derived peptide (Eisenhofer et al. 2004). In clinical practice, the only reliable criterion of malignancy is the presence of distant metastases.
Factors associated with better prognosis seem to include early diagnosis and excision of the primary tumor and, whenever possible, aggressive excision of any recurrence or soft-tissue metastases (Eisenhofer et al. 2004). In cases not suitable for surgery, pharmacological treatment with α-methyl-paratyrosine, an inhibitor of catecholamine synthesis, may improve quality of life but has no effect on tumor progression (Steinsapir et al. 1997). Conventional radiotherapy may provide effective palliation in patients with painful metastases. [131I]MIBG can be used as therapeutical option, given that the tumor shows uptake of MIBG. This may yield partial remission in 24–54% of patients and has even been reported to produce complete remission (Eisenhofer et al. 2004, Salmenkivi et al. 2004). If [131I]MIBG uptake is low, octreotide should be another therapeutic option, because several PHEOs express type 2 and 3 somatostatin receptors. However, the available data are rare and inconsistent (Eisenhofer et al. 2004). Another option is the application of high-dose [131I]MIBG (Rose et al. 2003). In case of no radionuclide uptake, chemotherapy should be administered. Given that there is no evidence-based chemotherapy protocol, the most widely used is the CVD protocol combining cyclophosphamide, vincristine, and dacarbazine (Averbuch et al. 1988). In rapidly progressive metastatic PHEOs, chemotherapy should be used as first-line therapy (Scholz et al. 2007). There are also reports for beneficial and additive effects of combined regime of chemotherapy and [131I]MIBG therapy (Shapiro et al. 2001, Sisson 2002). Radiofrequency ablation of hepatic and bone metastasis has shown promise in some selected cases (Pacak et al. 2001b). Promising new therapies might be developed such as inhibition of heat shock protein 90 and human telomerase reverse transcriptase or anti-angiogenic drugs (Eisenhofer et al. 2004).
Conclusions
It is widely accepted that a significant percentage of patients with sporadic PHEO and functional PGL may have germline mutations predisposing to the development of more generalized diseases. Several clinical characteristics, such as young age at onset, the presence of bilateral, extraadrenal, or multiple tumors or a malignant tumor should be seen as indications for genetic testing. The classical clinical picture, the biochemical testing, and the proper use of at least two imaging, anatomical and functional, techniques will confirm the diagnosis. The surgical removal of the tumor in sporadic cases will cure the patient, but in familial syndromes this has to be supplemented by a careful clinical follow-up for future recurrences. Therefore, in sporadic PHEOs, genetic testing will provide additional information necessary for a rational approach and management.
Acknowledgements
The authors declare that there is no conflict of interest that would prejudice the impartiality of this scientific work.
- © 2007 Society for Endocrinology
References
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
