This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Colao, A. Articles by Lombardi, G. Search for Related Content PubMed PubMed Citation Articles by Colao, A. Articles by Lombardi, G.

Systemic Complications of Acromegaly: Epidemiology, Pathogenesis, and Management

Department of Molecular and Clinical Endocrinology and Oncology (A.C., P.M., G.L.), "Federico II" University of Naples, 80131 Naples, Italy; and Department of Endocrinological and Metabolic Sciences and Center for Excellence for Biological Research (D.F.), University of Genova, 16132 Genova, Italy

Correspondence: Address all correspondence and requests for reprints to: Annamaria Colao, M.D., Department of Molecular and Clinical Endocrinology and Oncology, "Federico II" University, via S. Pansini 5, 80131 Napoli, Italy. E-mail: colao{at}unina.it.

	Abstract

Top Abstract I. Introduction II. The Complications at... III. The Metabolic Complications IV. The Complications at... V. The Neoplastic Complications VI. The Complications at... VII. Summary VIII. Conclusions References

 
This review focuses on the systemic complications of acromegaly. Mortality in this disease is increased mostly because of cardiovascular and respiratory diseases, although currently neoplastic complications have been questioned as a relevant cause of increased risk of death. Biventricular hypertrophy, occurring independently of hypertension and metabolic complications, is the most frequent cardiac complication. Diastolic and systolic dysfunction develops along with disease duration; and other cardiac disorders, such as arrhythmias, valve disease, hypertension, atherosclerosis, and endothelial dysfunction, are also common in acromegaly. Control of acromegaly by surgery or pharmacotherapy, especially somatostatin analogs, improves cardiovascular morbidity. Respiratory disorders, sleep apnea, and ventilatory dysfunction are also important contributors in increasing mortality and are beneficially advantaged by controlling GH and IGF-I hypersecretion. An increased risk of colonic polyps, which more frequently recur in patients not controlled after treatment, has been reported by several independent investigations, although malignancies in other organs have also been described, but less convincingly than at the gastrointestinal level. Finally, the most important cause of morbidity and functional disability of the disease is arthropathy, which can be reversed at an initial stage, but not if the disease is left untreated for several years.

I. Introduction

A. Epidemiology and causes of mortality in acromegaly

B. The clinical basis of increased mortality in acromegaly

C. The experimental basis for the GH/IGF-I effects at different body organs

D. Management of acromegaly

II. The Complications at the Cardiovascular System

A. Epidemiology

B. Pathogenesis

C. The acromegalic cardiomyopathy

D. Arrhythmias

E. Hypertension

F. Atherosclerosis and endothelial dysfunction

G. Effect of GH and IGF-I control on cardiovascular disease

III. The Metabolic Complications

A. Epidemiology

B. Pathogenesis

C. Effect of GH and IGF-I control on metabolic complications

IV. The Complications at the Respiratory System

A. Epidemiology

B. Pathogenesis

C. The sleep apnea syndrome

D. The respiratory dysfunction

E. Effect of GH and IGF-I control on respiratory disease

V. The Neoplastic Complications

A. Epidemiology

B. Pathogenesis

C. The gastrointestinal tract

D. Neck and lung tumors

E. Tumors of the reproductive system

F. Other tumors

VI. The Complications at the Skeletal System

A. Epidemiology

B. Pathogenesis

C. The acromegalic arthropathy

D. The carpal tunnel syndrome

E. Bone mass alterations

F. Effect of GH and IGF-I control on the skeletal system

VII. Summary

VIII. Conclusions

	I. Introduction

Top Abstract I. Introduction II. The Complications at... III. The Metabolic Complications IV. The Complications at... V. The Neoplastic Complications VI. The Complications at... VII. Summary VIII. Conclusions References

 
IN 1864, THE skull of a woman affected by prosopectasia (derived from the Greek words prosopon, face, and ektasis, stretching) was described by Verga (1) and added to the collection of the Anatomical Museum of Modena, Italy. During her lifetime, this patient suffered from typical somatic disfigurement, arrhythmias, and osteoarthropathy, although a postmortem examination revealed a giant pituitary (1). In 1881, Brigidi reported a description clinically consistent with acromegaly from the autopsy of the Italian actor Ghirlenzoni (2). This man had visceromegaly and enlarged hypertrophic pituitary. However, both Verga and Brigidi misinterpreted the pathogenesis of the syndrome, which was attributed to early menopause in the former and to primary bone disease in the latter case. Five years after Brigidi’s description, Pierre Marie (3) indicated with acromegaly his observation of two patients he had treated at the Salpetrière Hospital in Paris. At autopsy, Marie observed visceromegaly and enlarged pituitaries but was uncertain whether pituitary overgrowth was the cause etiology of such syndrome or whether it reflected the general process of organomegaly observed in these patients (3). Afterward, a progressively increasing number of similar descriptions were provided. Massalongo in 1892 and Benda in 1900 both indicated the cause of the disease as originating in the pituitary (2). However, it was only in 1909 that Harvey Cushing (4) reported the remission of clinical symptoms of acromegaly after partial hypophysectomy, thus indicating the etiology of the disease and its potential treatment as well. Acromegaly is known to be characterized by progressive somatic disfigurement and a wide range of systemic manifestations (5, 6). At diagnosis, patients generally exhibit coarsened facial features, exaggerated growth of hands and feet, and soft tissue hypertrophy (Table 1). Other characteristics may include hyperhidrosis, goiter, osteoarthritis, carpal tunnel syndrome, fatigue, visual abnormalities, increased number of skin tags, colon polyps, sleep apnea and daytime somnolence, reproductive disorders, and cardiovascular disease, which most commonly includes cardiac hypertrophy, hypertension, and moderate arrhythmias, although congestive heart failure occurs more rarely (5, 6, 7). Improvement of surgical procedures, radiotherapy tools, and the availability of pharmacological compounds active on somatotroph pituitary cells greatly changed the approach to this disease that is an extraordinary model to investigate the pathophysiology of GH and IGF-I actions on virtually all body organs and systems; several systemic consequences developing in the course of undiagnosed GH and IGF-I excess may remain undiagnosed for a long time.

A. Epidemiology and causes of mortality in acromegaly
Acromegaly has an estimated annual incidence of three to four cases per million population and a current estimated prevalence of 40 cases per million population, which can be reportedly as high as 90 cases per million population (8, 9, 10). In most cases, chronic GH hypersecretion is caused by a benign pituitary adenoma (5). Due to its indolent and insidious nature, the diagnosis of acromegaly is usually delayed for a variable number of years (7). Retrospective epidemiological studies showed that diagnosis can be preceded by approximately four to 10 or more years of active disease, and it is likely that today the diagnosis is made earlier than before because patients present much less with visual field defects (11). Analysis of determinants of mortality outcome indicates that approximately 60% of acromegalic patients die from cardiovascular disease, 25% from respiratory disease, and 15% from malignancies (12, 13, 14, 15, 16, 17, 18, 19). The nadir GH value likely constitutes the most predictive survival index, regardless of the death cause (13, 15, 16, 17, 18). There is compelling evidence indicating that control of GH levels and/or IGF-I levels normalized for age is associated with improvement of adverse mortality rates, independent of the type of associated complications (17). Suppression of GH below 5 mU/liter (<2.5 µg/liter) had been shown already to portend a favorable mortality outcome (19, 20). In fact, it has been suggested that overall mortality in patients with acromegaly is correlated with the degree of GH control (18); mortality rates for cancer can be stratified according to posttreatment GH levels (18); and, if GH secretion is controlled, mortality rates become similar to those recorded in the nonacromegalic population (13, 18). However, mortality in patients with cardiac disease at the time of diagnosis occurs within 15 yr in almost 100% of cases, and only 20% of patients with diabetes and acromegaly will survive 20 yr (16). High GH levels, hypertension, and heart disease constitute the major negative survival determinants in acromegaly (19), whereas symptoms duration and other factors, including uncontrolled diabetes and/or dyslipidemia and cancer, account less for mortality (Fig. 1). Thus, control of GH hypersecretion, hypertension, and heart disease is relevant to improve the ultimate mortality rates.

View larger version (21K): [in this window] [in a new window]   FIG. 1. Depiction of mortality determinants in patients with acromegaly. The x-axis reflects the P value (log) as calculated from published retrospective reports. [From S. Melmed: J Clin Endocrinol Metab 86:2929–2934, 2001 (19 ). Permission granted by The Endocrine Society.]

C. The experimental basis for the GH/IGF-I effects at different body organs
Nearly one century after Verga’s original documentation (1), Salmon and Daughaday’s findings (26) set the basis for the somatomedin hypothesis with the identification of liver-derived serum factor-mediating GH actions on peripheral tissues. The same serum factor was sequentially termed sulfation factor, nonsuppressible insulin-like activity, multiplication-stimulation activity, and eventually IGF-I in independent investigations (reviewed in Refs. 27 and 28). The concept that GH operated on peripheral tissues through endocrine IGF-I was successively revised in the early 1980s, when IGF-I was shown to be expressed ubiquitously in several body tissues and its local expression could be modulated by exogenous GH administration to animals (27, 28). Subsequent studies in preadipocytes and chondrocytes attempted at discriminating between GH and IGF-I effects in the periphery and demonstrated that GH promoted cellular differentiation, whereas IGF-I preferentially induced clonal expansion, thus introducing the dual effector theory (29). However, the congruity of the aforementioned theories has been recently questioned. GH has been found also to modulate the expression of various growth factors and their receptors in several tissues, in addition to IGF-I. GH can stimulate directly the proliferation of chondrocytes in the germinal zone of the growth plate (30), and it promotes the activation of immediate early gene (i.e., transcription of early growth response factor-1) in 3T3-F442A preadipocytes (31). A demonstration that endocrine GH had local effects on cells originated from studies in mice carrying null mutations for the IGF-I gene that showed a normal proliferation of growth plate chondrocytes despite a 35% decrease in longitudinal bone growth (32). To explain this finding, the authors speculated that the lack of IGF-I stimulus had likely been replaced by endogenous GH or, alternatively, by local IGF-II (32). Recent investigations are now questioning whether liver IGF-I production actually plays an essential role in postnatal growth and development. Studies in conditional IGF-I knockout mouse models obtained with the Cre/loxP recombination system have, in fact, demonstrated that abrogation of liver IGF-I expression blunts 75% of endocrine IGF-I secretion, increases GH levels in serum likely due to feedback mechanisms, and preserves postnatal growth and development; as a result, autocrine/paracrine IGF-I production seems to be sufficient for normal growth and development, although a direct effect of GH on nonhepatic tissues cannot be excluded (33, 34).

GH action is achieved via its interaction on the cell surface with its receptor (GH-R) belonging to the family of the cytokine receptors (35), whereas IGF-I belongs to a system of proteins encompassing IGF-II, the type-I and -II IGF receptors (IGF-IR and -IIR), six IGF binding proteins (IGFBPs), three families of IGFBP proteases, and nine growth-mediating factors that share structural similarities in the N-terminus of the IGFBPs and have therefore also been termed IGFBP-related proteins (Fig. 2) (28). IGF-I circulates in serum bound to IGFBP-3 and the acid-labile subunit (ALS) in a 150-kDa ternary complex, which works both as a reservoir and regulator of IGF-I biodistribution to peripheral tissues/organs. Due to the molecular weight of ALS, IGF-I is unable to cross the endothelium unless it is released from the ternary complex after the proteolysis of IGFBP-3. IGFBP-3 proteolysis generates fragments with a lower IGF-I binding affinity, and this is key in the regulation of local IGF-I bioavailability. Studies in hypophysectomized rats have clearly demonstrated that all three peptides of the ternary complex are synthesized in the liver in a GH-dependent fashion, although a GH-responsive element has only been identified in the ALS promoter (36). In the liver, hepatocytes are the primary source for IGF-I and ALS production, whereas IGFBP-3 is synthesized preferentially in the portal venous and sinusoidal endothelium, as determined by the expression of mRNA by in situ hybridization histochemistry (37). Due to their GH dependence, the impairment of total and free IGF-I, IGFBP-3, and ALS secretion is used for diagnostic purposes in acromegaly and GH deficiency (38, 39). Although it has been shown that both IGF-I and IGFBP-3 elicit multiple cellular effects both dependently and independently of each other, ALS does not seem to possess intrinsic biological activity. The IGF-I cellular effects are achieved through the interaction of IGF-I with the IGF-IR, a heterotetrameric protein that is evolutionarily and structurally similar to the insulin receptor in its - and ß-subunits (40). The intracellular domains of the ß-subunits of IGF-IR possess tyrosine kinase activity and tyrosine residues that are phosphorylated upon receptor activation. The consequence of receptor phosphorylation is the downstream recruitment of signaling proteins including the IRS and Shc families of proteins, which in turn activate the MAPK and phosphatidylinositol 3-kinase pathways, as well as additional downstream signaling proteins leading to gene transcription (27). Depending on the preponderance of intracellular IRS or Shc proteins, IGF-I will promote either cellular proliferation/transformation or differentiation (41). Noteworthy, the primary role of IGF-I is the regulation of postnatal growth and the mediation of the growth-promoting effects of GH. The overexpression of bovine, murine, or rat GH stimulates body growth and increases IGF-I levels in transgenic mice (42). Similar body growth promoting effects are produced by human IGF-I overexpression in the mice liver (43). In contrast, IGF-I null mice suffer from impairment of fetal growth and development, whereas perinatal viability is markedly decreased (44). The physiological actions of IGF-I also encompass increased muscle protein synthesis (45), in vivo and in vitro stimulation of muscle cell differentiation (40), enhanced glucose uptake in peripheral tissues (45), bone growth and maturation (46), oligodendrocyte survival, and neuronal differentiation (47). These effects require a functionally active IGF-IR, which is essential for normal embryonic and fetal growth, modulation of cellular apoptosis, and the growth, proliferation, and migration of normal and tumoral cells (48).

View larger version (18K): [in this window] [in a new window]   FIG. 2. Representation of the IGF system comprising IGFs, IGFBP-1 through 6, IGFBP-related proteins, IGF and IGFBP cell membrane receptors, and additional translocation mechanisms for IGFBP-3 to the nucleus.

D. Management of acromegaly
An extensive discussion on the current approaches for an integrated and modern management of acromegaly lies beyond the scope of this review. However, because most of the subsequent parts will refer to the partial or total recovery of systemic complications of acromegaly after controlling GH/IGF-I hypersecretion, the most recent advances in the management of acromegaly will be summarized here.

The optimal treatment for acromegaly should be able to remove the tumor with resolution of its mass effects, preserve the normal residual pituitary function, prevent recurrences, restore normal GH and IGF-I secretion, relieve symptoms directly caused by GH excess, and, possibly, prevent progressive disfigurement, bone expansion, osteoarthritis, cardiomyopathy, hypertension, insulin resistance, diabetes mellitus, and lipid abnormalities, thus reversing the unfavorable long-term outcome (6). The treatment options currently available for such ambitious goals include surgery, irradiation, and pharmacological suppression of GH levels by somatostatin analogs or dopamine agonists, and/or by functional blockade of the GH-R by modified GH analog (53), which has provided promising results (54, 55). Cure criteria to evaluate the efficacy of all treatment approaches, except for the GH-antagonist, are mean integrated 24-h GH levels less than 2.5 µg/liter or no more than 1 µg/liter after glucose load, together with circulating IGF-I levels normalized for age and gender (20). Clearly, other pituitary hormone deficiencies that may be caused by the tumor and/or by its treatment must be treated as in other types of pituitary tumors.

Transsphenoidal adenomectomy remains a milestone treatment for GH-secreting tumors, but should be performed only in experienced centers. Moreover, it has been reported recently that the outcome of surgery when only one surgeon operates on patients is higher than that observed when more than one surgeon operates (56, 57). The outcome of surgery is successful in most microadenomas and enclosed macroadenomas, but it remains disappointing in larger adenomas and negligible in invasive tumors (6, 16, 56, 58). Among 224 consecutive patients, endocrine remission occurred in 72% of microadenomas, 50% of macroadenomas, and only 17% of giant adenomas (59). Normalized IGF-I levels with GH levels below 3 µg/liter (60), or 2.5 µg/liter (61), were achieved in 59% and 42% of unselected patients, respectively. In microadenomas, success of surgery was obtained in 61% of patients (61). Surgery relieves the compression on adjacent structures such as optic chiasm and ventricles, and it only rarely causes complications. In fact, in a study including questionnaires regarding 14 specific complications of transsphenoidal surgery mailed to 3172 neurosurgeons, it was reported that the mean operative mortality rate was low (0.9%), anterior pituitary insufficiency and diabetes insipidus were the most common complications (19.4% and 17.8%, respectively), whereas cerebrospinal fluid fistulas were found in a low number of cases (3.9%) and other complications, such as carotid artery injuries, hypothalamic injuries, loss of vision, and meningitis, occurred in 1–2% of cases (62). In another study (63), pituitary failure after surgery was confirmed to be the most frequent complication occurring in up to 30% of patients with macroadenomas. A controversy exists on the usefulness of a short preoperative drug treatment to improve surgical outcome (for review, see Refs. 6 and 64); we observed that a 6-month treatment with octreotide preoperatively improved metabolic and hemodynamic parameters and reduced the duration of hospital stay in acromegalic patients (65). Recently, to provide a technical refinement in pituitary surgery, a one-nostril endoscopic endonasal transsphenoidal procedure has been proposed to reduce the damage to nose and sphenoid sinus and to improve the management of the pituitary region (66). Interestingly, postsurgical complications using this technique are likely to be low compared with classical microscopical approach (67). In fact, among 146 consecutively treated patients who underwent an endoscopic endonasal transsphenoidal approach to the sellar region for resection of pituitary adenomas between January 1997 and July 2001, complications (divided into groups, i.e., nasofacial, sphenoid sinus, sella turcica, supra or parasellar, and endocrine) were decreased in their incidence compared with large historical series of the traditional microsurgical transsphenoidal approach (67). The overview inside the anatomy facilitated by the endoscope and the consequent decreased surgical trauma can be taken as explanation for such positive findings.

Irradiation of GH-secreting pituitary tumor should be reserved to patients in whom surgery is contraindicated or unsuccessful and medical treatment fails to control persistent hormone hypersecretion due to a remnant tumor (68). Multiple methods for delivery of radiation are currently used, including external radiation, proton beam, -particles, and interstitial radiotherapy. All seem to induce similar cure rates. The greatest fall in GH levels occurs within the first 2 yr, followed by a gradual decline thereafter for 10 yr (69). Earlier observations that pituitary irradiation normalized IGF-I concentrations only in a minority of patients (70, 71) have not been confirmed by later studies reporting higher cure rates (72, 73). Furthermore, the occurrence of severe complications, such as cranial nerve palsies, optic neuritis, impaired memory, lethargy, and tissue necrosis has decreased with modern techniques. However, damage of the normal hypothalamic-pituitary region results in hypopituitarism in more than half of patients within 10 yr (6, 63, 68, 69). Gamma-knife radiosurgery has been used as adjuvant treatment for pituitary adenomas in selected cases with promising results (74), but definitive data are still lacking.

Bromocriptine, the first and likely still the most widely used dopamine agonist in acromegaly, lowered GH levels below 10 µg/liter and 5 µg/liter in 50% and 10–20%, respectively, of over 500 patients with acromegaly included in a metaanalysis from 28 published series (75). It produced improvement of symptoms of acromegaly in up to 70% of the patients, but tumor shrinkage was rare; very high doses (10–20 mg/d) are generally required, and side effects are common (76). Variable results have been reported recently using cabergoline, a selective D₂ receptor agonist more potent than bromocriptine (77, 78, 79). Disease control is more likely achieved in patients with mixed prolactin (PRL)/GH-secreting adenomas than in the pure GH-secreting ones, and in patients with lower GH and IGF-I levels before treatment than in those with more aggressive disease (6).

Somatostatin analogs are, at present, the most widely used drugs to control acromegaly. Octreotide is an octapeptide, displaying a high affinity for somatostatin receptor subtypes 2 and 5 and a faint affinity for subtype 3, which has been largely used in acromegaly with excellent results (80). Octreotide given sc at a dose of 100–250 µg every 8 h for 6 months reduced GH levels below 5 µg/liter in 53% and normalized IGF-I levels in 68% of 115 patients enrolled in a multicenter placebo-controlled study (81). A metaanalysis of 466 patients treated worldwide showed that octreotide suppressed GH levels below 2.5 µg/liter in 29.2%, normalized IGF-I levels in 39.9%, and reduced tumor size (>20% of reduction in maximal diameter) in 38.6% of patients; the GH-lowering effect was related to initial GH values (6). Most clinical signs and symptoms of acromegaly, such as sweating, soft tissue swelling, fatigue, and headache, are generally relieved after the administration of the first doses of octreotide. Twenty percent of patients develop gallbladder abnormalities (biliary sediment/sludge, microlithiasis, or gallstones), but morbidity is negligible, and treatment with ursodehoxicolic acid can be performed in those patients with symptomatic gallstones (80). Lanreotide is another analog, showing a binding profile comparable to octreotide, and has a similar efficacy in suppressing GH and IGF-I levels (6, 68). Depot preparation of lanreotide and octreotide long-acting repeatable (LAR) have further improved the therapeutic success of sc octreotide (68). In particular, we showed tumor shrinkage, graded from mild to notable, in 80% of 15 newly diagnosed patients treated for 12 months with octreotide-LAR, suggesting its potential application as primary therapy in invasive adenomas (82). Our data have been subsequently confirmed by the results of a multicenter prospective study showing tumor shrinkage in all 27 newly diagnosed patients with acromegaly with median tumor volume reduction of 49% in microadenomas and 43% in macroadenomas (83). However, because a large variability in tumor shrinkage has been reported, it is still hard to estimate tumor size response to slow-release somatostatin analogs. Both depot formulations are well tolerated; the mild-to-moderate side effects experienced by up to 50% of the patients have short duration and often subside with treatment continuation (6, 68, 80).

As anticipated, the newest drug for treating acromegaly bases its efficacy on blocking the activity of the GH-R, thereby inhibiting the synthesis of IGF-I (53). In a placebo-controlled study, there was a significant dose-dependent fall in serum IGF-I in three groups treated with the GH-antagonist compared with placebo-treated patients, and 90% of patients treated with the highest dose (20 mg) achieved normal IGF-I levels for age (54). In line with IGF-I decrease, IGFBP-3 similarly decreased, and patients experienced an improvement of physical well-being and clinical signs. These results persisted until 24 months (55). Even if data on this new drug are still scant, the GH-antagonist seems to be well tolerated, except for rare cases of increased hepatic transaminase levels; two of 133 patients had increased tumor mass (54, 55), one of them being stabilized after a combined treatment with the GH-antagonist plus octreotide (84). Currently, the use of GH-antagonist is permitted in the United States and is still experimental in the rest of the world, but the drug should be available for treatment of acromegaly by the beginning of 2004.

	II. The Complications at the Cardiovascular System

Top Abstract I. Introduction II. The Complications at... III. The Metabolic Complications IV. The Complications at... V. The Neoplastic Complications VI. The Complications at... VII. Summary VIII. Conclusions References

 
A. Epidemiology
GH and IGF-I elicit primary regulatory activities both in developing heart growth and in maintaining its structure (21). In the theoretical absence of other cardiac diseases, the involvement of the heart in acromegaly defines the acromegalic cardiomyopathy, which was first described at the end of the 19th century (85). Although the prevalence of the acromegalic cardiomyopathy has not been investigated in detail, its most common feature is considered to be a concentric biventricular hypertrophy (21, 86, 87, 88, 89, 90). Cardiac walls are generally thickened, but cardiac chambers are rarely enlarged due to the relative increase of cardiac myocyte width for the parallel assembling of new contractile-protein units (91). Aging and long duration of GH/IGF-I excess are main determinants of cardiac derangement; results collected in vivo and postmortem showed a prevalence of cardiac hypertrophy higher than 90% in patients with long disease duration (92, 93). However, more recent surveys demonstrated that structural changes of the heart can even occur in patients shortly exposed to GH hypersecretion (94, 95, 96), and 20% of normotensive patients younger than 30 yr develop cardiac hypertrophy (96). In particular, we recently reported that the left ventricular mass index was approximately 30% higher in 25 acromegalic patients below age 40 yr than in 25 age-matched control subjects (Fig. 3); 60% of the patients had clear-cut left ventricular hypertrophy (97). Characteristically, the cardiac hypertrophy of acromegaly occurs in the absence of hypertension that is present in approximately one third of patients (see Section II.C) and is further aggravated by hypertension and glucose abnormalities. In our analysis (22), 100% of patients with hypertension and diabetes had cardiac abnormalities at echocardiography (Fig. 4). The acromegalic cardiomyopathy develops after three steps: 1) in the early phase, and thus mainly in young patients with a short disease duration, there is initial cardiac hypertrophy, high heart rate, and increased systolic output altogether configuring the hyperkinetic syndrome (98); 2) in the middle phase, hypertrophy becomes more evident, signs of diastolic dysfunction appear, and insufficient systolic function on effort can be documented; and 3) in the end-stage of untreated disease, cardiac abnormalities may include systolic dysfunction at rest and heart failure with signs of dilative cardiomyopathy (21, 99). It should be considered, however, that longer acromegaly duration is generally accompanied by an older age, and it is well known that aging is accompanied by significant cardiovascular modifications, both structural and functional (100). In the nonacromegalic population, aging is associated with a slight degree of left ventricular hypertrophy and decrease, even modest, of resting heart rate and early filling rate, whereas end-diastolic and end-systolic dimensions, stroke volume, and ejection fraction are largely unchanged. It should be noted that in acromegaly the prevalence of left ventricular hypertrophy is predominant but, besides hypertrophy, the majority of the patients at diagnosis have a normal (55–78%) left ventricular ejection fraction in resting conditions (23).

View larger version (21K): [in this window] [in a new window]   FIG. 3. Effect of age on the left ventricular mass index (LVMi, left) and response of left ejection fraction at peak exercise (LVEF, right) in our series. Acromegalic patients and controls were grouped according to age [<40 yr (n = 32), 40–60 yr (n = 33), and >60 yr (n = 22)]. Data are derived from our own experience and include data reported in Refs. 97 , 191 , and 199 . *, P < 0.001 vs. controls; **, P < 0.001 vs. before treatment that consisted of octreotide-LAR at a dose of 20–40 mg/month.

View larger version (47K): [in this window] [in a new window]   FIG. 4. Prevalence of left ventricular hypertrophy, impaired diastolic filling, and inadequate ejection fraction at rest in 130 patients with acromegaly studied at their diagnosis and grouped on the basis of the absence of hypertension and glucose tolerance abnormalities, absence of hypertension and IGT, absence of hypertension and diabetes mellitus, hypertension and glucose tolerance abnormalities, hypertension and IGT, hypertension and diabetes mellitus. [From A. Colao et al.: J Clin Endocrinol Metab 85:193–199, 2000 (23 ). Permission granted by The Endocrine Society.]

B. Pathogenesis
The mechanisms of GH and IGF-I action on the heart have been reviewed recently (21, 107). The effects of GH, IGF, and their binding peptides on the cardiovascular system are both direct, via endocrine, autocrine, and paracrine mechanisms, and indirect because they cause increased cardiovascular risk and hypertension. GH and IGF-I receptors are expressed in cardiomyocytes (108, 109), and IGF-I mRNA is present in the epicardium and the coronary vessel of the human fetus (110). Neonatal rat ventricle tissue preferentially expresses IGF-II mRNA transcripts and contains both IGF-IR and IGR-IIR (111), although cardiac IGF-IR expression is partly blunted in the adult rat (109, 112). Interestingly, IGF-I immunoreactivity is reportedly increased in the inner layers of the left ventricle (113), where both tension and wall stress are high, and gradually decreases toward the epicardial surface (114). IGF-I expression accompanies the development of left ventricular hypertrophy (113, 115); in rat myocardium, IGF-I mRNA is increased after pressure overload secondary to banding of ascending aorta, aorto-caval shunt, myocardial infarction, experimental renal, or pulmonary hypertension (113, 116, 117, 118). However, the role of IGF-I in heart development during the prenatal life is still unclear. In fact, newborns of mice knocked-out for the IGF-I gene show a reduced body size compared with control littermates, but heart size is generally unaffected (119). Similarly, when the IGF-I synthesis in the liver is abolished by the Cre/loxP recombination system, no negative effect on the postnatal cardiac size is found (34). In the GH-R knockout mouse, there was no change in heart weight (120), although GH secretion rate has a prominent role in the postnatal cardiac development. Hypophysectomy in rats induces a decreased cardiac expression of IGF-I mRNA, which can be restored by exogenous GH administration (121). GH administration in hypophysectomized rats with moderate myocardial infarction does not improve ventricular function (122), whereas GH-secreting tumors implanted in rats determine cardiac hypertrophy, enhance the contractile performance, and produce the elongation of the action potential of cardiac fibers (91, 123). Exogenous GH and IGF-I administration in normal adult rats induces a hypertrophic response of the heart without developing significant fibrosis (124). Furthermore, IGF-I increases the intracellular calcium content and enhances the calcium sensitivity of myofilaments in cardiomyocytes (125). Cardiomyocyte stimulation induced by GH and IGF-I is associated with a low-energy conformational status, mediated by myosin pheno-conversion from the isoform V3 to a low ATPase activity isoform (126). GH, either directly (127) or via IGF-I, increases myocardial hypertrophy (127) and increases myocardial contractility in animal models of chronic GH excess (128) and in cardiomyocytes from neonatal rats (129), likely via an increased calcium responsiveness of myofilaments (130).

C. The acromegalic cardiomyopathy
Although new sophisticated methods are currently available to study the acromegalic cardiomyopathy, echocardiography still remains the most used method (131). Clear-cut left ventricular hypertrophy is found in most patients at diagnosis, overall in those with long disease history (21), and interstitial fibrosis constitutes the main abnormality at histology (91, 93, 123). Subsequently, gradual impairment of heart architecture by increased extracellular collagen deposition, myofibrillar derangement, areas of monocyte necrosis, and lympho-mononuclear infiltration occurs, thus configuring a pattern of myocarditis (92, 93). Increase of apoptosis in cardiomyocytes and interstitial fibroblasts, inversely correlated to the output rate, was found in biopsied cardiac tissue obtained during heart catheterization in acromegalic patients (132). In a survey performed in our department including 200 patients undergoing echocardiography at diagnosis, left ventricular hypertrophy was found in 120 patients (60%); the left ventricular mass index significantly increased from young (<30 yr) to elderly (>60 yr) patients (Fig. 5). Accordingly, the prevalence of left ventricular hypertrophy was higher in patients older than 50 yr (74.3%) than in younger patients (57% in patients aged 31–50 yr and 35% in those aged <30 yr; our unpublished data). As already mentioned, left ventricular hypertrophy is not negligible in young patients with a presumed short duration of acromegaly, confirming previous data of other groups (95, 96) as well as from ours (97). This suggests that cardiac hypertrophy is an early event in acromegaly, which worsens proportionately with the duration of disease activity. It is known that arterial hypertension is likely the most important factor aggravating cardiac hypertrophy and has higher prevalence in aged patients (see Section II.E). In another study including a large series of patients (23), we observed that hypertension significantly increased the impact of cardiac hypertrophy, therein documented in 51% of cases. The prevalence of hypertrophy was higher in hypertensive patients (Fig. 4), and the multistep regression analysis showed that the diastolic blood pressure was the best predictive factor of cardiac hypertrophy (23). It should be mentioned that patients with hypertension and diabetes had an older age than those with uncomplicated acromegaly. Because aging in nonacromegalic subjects is characterized by a slight increase in left ventricular hypertrophy (100), it is likely that in acromegaly this phenomenon is emphasized. It should be stated, however, that it is currently unknown whether aging has independent negative effects on the heart in acromegaly, because there are no controlled studies in the elderly patients population. In our series, however, patients older than 60 yr had a significantly higher left ventricular mass than age- and sex-matched controls, who indeed had an increased mass compared with young controls (Fig. 3). It is also unknown whether there are gender differences in the prevalence and severity of the acromegalic cardiomyopathy. Gender difference is well known to occur in GH and IGF-I secretion both in healthy subjects (133) and in acromegaly (134, 135). Reviewing our experience in 200 patients with acromegaly, we did not find any difference in the prevalence of left ventricular hypertrophy between women (63.6%) and men (58%); similarly, the prevalence of hypertension and diabetes was similar in both sexes (our unpublished data). A minor but relevant complication that, similar to hypertension and diabetes, may further complicate acromegalic cardiomyopathy is thyrotoxicosis (136), which appeared to primarily affect systolic function.

View larger version (51K): [in this window] [in a new window]   FIG. 5. Results of the echocardiography study performed at diagnosis in 200 patients with acromegaly studied at the Department of Molecular and Clinical Endocrinology and Oncology, University "Federico II" of Naples. Data are shown according with patients’ age divided into decades. The left ventricular mass index (LVMi, top left) significantly increases with aging, whereas diastolic filling, measured as early to late mitral flow velocity (E/A, top right), and left ventricular ejection fraction at rest (LVEF, bottom left) significantly decrease with aging. As a consequence, the prevalence of left ventricular hypertrophy (LVH), inadequate diastolic filling, and systolic performance increases with aging.

View larger version (21K): [in this window] [in a new window]   FIG. 6. Linear correlation analysis in patients (•) and controls () between age and left ventricular ejection fraction at rest (top) and at peak exercise (bottom). [From A. Colao et al.: J Clin Endocrinol Metab 84:1518–1523, 1999 (137 ). Permission granted by The Endocrine Society.]

View larger version (13K): [in this window] [in a new window]   FIG. 7. Linear correlation analysis between the estimated disease duration and the diastolic filling, measured as the ratio between the peak filling rate and the peak ejection rate. [Data were modified from A. Colao et al.: J Clin Endocrinol Metab 84:1518–1523, 1999 (137 ). Permission granted by The Endocrine Society.]

E. Hypertension
Arterial hypertension is considered one of the most relevant negative prognostic factors for mortality in acromegaly. However, both epidemiology and pathophysiological mechanisms are far from totally clear. Hypertension is reported to affect approximately one third of patients with acromegaly, but only a few studies have estimated the prevalence of hypertension using the 24-h pressure Holter recording (147, 148, 149, 150), and no control populations were included in these studies. The only study including a control group (151) revealed an increased prevalence of hypertension only in patients with familiarity for hypertension. In a survey study of 200 patients with acromegaly studied at diagnosis, we have found hypertension (based on a diastolic blood pressure > 90 mm Hg) in 40%, compared with 8% of controls; there was no gender difference in the patients or in controls (our unpublished data).

One mechanism likely contributing to inducing hypertension in acromegaly is represented by the increased plasma volume (152, 153). The evidence of an increase in the total exchangeable sodium pool in normotensive (154, 155, 156, 157) and hypertensive acromegalic patients (151) further supports this hypothesis, because a direct relationship was demonstrated between total exchangeable sodium and blood pressure values (156, 157). Whether the sodium exchange pump is also involved is unclear, because either a reduction (158) or an increase (159) of its activity has been reported. Plasma volume and total exchangeable sodium were initially supposed to follow variations of aldosterone secretion. However, the levels of aldosterone, and its precursors corticosterone and 11-deoxy-corticosterone as well, were found to be normal in patients with acromegaly without (152, 157, 160, 161) or with (151) hypertension. Aldosterone levels were not correlated with disease activity unless overt heart failure was present (152). Similarly the aldosterone response to stimuli, such as posture or saline infusion, gave contradictory responses (152, 160, 162). Both the atrial natriuretic peptide (ANP) and the renin-angiotensin system have been claimed as potential causes to explain hypertension in acromegaly without success. In fact, basal ANP levels were found to be normal (160), not correlated with disease activity (152), not increased after saline infusion (160) or increased as in controls (152). Similarly, basal PRA was normal in most studies (152, 157, 160, 161), not correlated with disease activity (152) or hypertension (151), and reduced (155, 161) or inappropriately reduced to hypernatremia (155, 156); the response of the renin activity to stimuli was also unclear (152, 153, 157, 160, 161). The adrenergic system has also been investigated, but no clear evidence for its involvement emerged from different studies. Plasma epinephrine and/or norepinephrine levels were normal (157, 163, 164) or increased (165), and urinary catecholamines were also normal (157, 163, 166) or increased (167). Bondanelli et al. (162) reported that plasma catecholamines were normal but without the physiological circadian rhythm (Fig. 8) that reappeared after successful surgery (Fig. 8). No difference between patients with acromegaly and controls was found in catecholamine levels both basally and after hyperinsulinemic clamp (168). No change of catecholamine levels was found after bromocriptine or lysuride treatment (164), whereas norepinephrine levels were reduced in another study (167).

View larger version (34K): [in this window] [in a new window]   FIG. 8. Plasma norepinephrin (NE) and epinephrin (E) concentrations in blood obtained at 1-h intervals for 24 h from normal subjects and acromegalic patients before transsphenoidal surgery (A) and from acromegalic patients in remission or active disease after transsphenoidal surgery (B). [From M. Bondanelli et al.: J Clin Endocrinol Metab 84:2458–2467, 1999 (162 ). Permission granted by The Endocrine Society.]

View larger version (21K): [in this window] [in a new window]   FIG. 9. Linear correlation analysis between fasting insulin levels and diastolic blood pressure in 200 patients with acromegaly () and 100 controls () studied at the Department of Molecular and Clinical Endocrinology and Oncology, University "Federico II" of Naples.

View larger version (46K): [in this window] [in a new window]   FIG. 10. Capillaroscopy in one control (A) and one patient with acromegaly (B). Note the remarkable meandering of several capillaries. [From F. Schiavon et al.: J Clin Endocrinol Metab 84:3151–3155, 1999 (175 ). Permission granted by The Endocrine Society.]

F. Atherosclerosis and endothelial dysfunction
Few data are currently available on the vascular involvement in acromegaly. Cardiac output was shown to be heterogeneously distributed with lower regional brachial artery blood flow and increased local resistance (172). The study of the peripheral microcirculation (Fig. 10) showed a significantly lower capillary number and length and a significantly higher number of tortuous loops and meandering capillaries in patients with acromegaly than in controls (175). The capillaroscopic alterations were still observed in a smaller group of patients not bearing diabetes and hypertension (175). Increase of the carotid intima-media thickness (IMT) was observed in active as well as cured patients with acromegaly, but the prevalence of well-defined atherosclerotic plaques was not higher than in control subjects (177). It should be noted that in the group of cured acromegalic patients insulin, cholesterol, and fibrinogen levels were still slightly higher than in controls (177). The presence of still elevated insulin levels in patients cured from acromegaly can be the underlying factor able to maintain an increased IMT, because insulin levels are known to be directly correlated with IMT (178). In another series of patients with acromegaly, prospectively studied before and after 6 months of treatment with lanreotide, an increase of IMT at the level of common carotid arteries was similarly observed, and 29 of them had abnormal IMT levels (179). Only a mild increase of carotid IMT was conversely reported by Kasayama et al. (180); however, because plasma IGF-I concentration was significantly higher and the prevalence of hypertension was significantly lower in patients without than in those with atherosclerotic changes, the authors concluded that increased concentration of IGF-I might be involved in the lack of susceptibility to atherosclerosis in some acromegalic patients (180). We could not find, however, any correlation between IGF-I levels and carotid IMT (177, 179).

Laser Doppler flowmetry has also confirmed endothelial dysfunction at the hand cutaneous circulation. Vascular smooth cell ability to produce skin vasodilatation was normal, but endothelium-dependent vasodilatation was impaired and sympathetic-mediated vasoconstrictive response was increased in normotensive acromegalic patients (181). Very recently, we demonstrated that the increased IMT in patients with acromegaly mainly depends on concomitant risk factors, because there was no difference between patients with active or cured acromegaly and their controls matched for hypertension, diabetes, or dyslipidemia (182). Interestingly, the endothelium-dependent vasodilatation, measured at the brachial artery level, was impaired in patients with active acromegaly more than that expected on the basis of classical risk factors (Fig. 11); this allowed us to hypothesize a direct negative effect of GH and IGF-I hypersecretion on endothelial function (182). Clearly, the existence of other negative factors, such as glucose intolerance, dyslipidemia, and smoking habitus, further impairs vascular relaxation.

View larger version (9K): [in this window] [in a new window]   FIG. 11. Boxes showing median values () and interquartile ranges of flow-mediated dilatation (FMD) in the study population. Active, Active acromegalics; Active MC, active acromegalic-matched controls; Cured, cured acromegalics; Cured MC, cured acromegalic-matched controls. *, Lower than in healthy subjects (P < 0.01); , lower than in active MC (P < 0.01); #, lower than in cured (P < 0.01); , lower than in healthy subjects (P < 0.05); lower than in cured MC (P < 0.05). [From G. Brevetti et al.: J Clin Endocrinol Metab 87:3174–3179, 2002 (182 ). Permission granted by The Endocrine Society.]