Annamaria Colao, Department of Molecular and Clinical Endocrinology and Oncology, University of Naples Federico II, Via S. Pansini 5; 80123 Naples, Italy. Tel.: +39 081 7462132; Fax: +39 081 5465443 or +39 081 7463668; E-mail: firstname.lastname@example.org
Background GH and IGF-I affect cardiac structure and performance. In the general population, low IGF-I has been associated with higher prevalence of ischaemic heart disease and mortality. Both in GH deficiency (GHD) and excess life expectancy has been reported to be reduced because of cardiovascular disease.
Objective To review the role of the GH–IGF-I system on the cardiovascular system.
Results Recent epidemiological evidence suggests that serum IGF-I levels in the low-normal range are associated with increased risk of acute myocardial infarction, ischaemic heart disease, coronary and carotid artery atherosclerosis and stroke. This confirms previous findings in patients with acromegaly or with GH-deficiency showing cardiovascular impairment.
Patients with either childhood- or adulthood-onset GHD have cardiovascular abnormalities such as reduced cardiac mass, diastolic filling and left ventricular response at peak exercise, increased intima-media thickness and endothelial dysfunction. These abnormalities can be reversed, at least partially, after GH replacement therapy. In contrast, in acromegaly chronic GH and IGF-I excess causes a specific cardiomyopathy: concentric cardiac hypertrophy (in more than two-thirds of the patients at diagnosis) associated to diastolic dysfunction is the most common finding. In later stages, impaired systolic function ending in heart failure can occur, if GH/IGF-I excess is not controlled. Abnormalities of cardiac rhythm and of cardiac valves can also occur. Successful control of acromegaly is accompanied by decrease of the left ventricular mass and improvement of cardiac function.
Conclusion The cardiovascular system is a target organ for GH and IGF-I. Subtle dysfunction in the GH–IGF-I axis are correlated with increased prevalence of ischaemic heart disease. Acromegaly and GHD are associated with several abnormalities of the cardiovascular system and control of GH/IGF-I secretion reverses (or at least stops) cardiovascular abnormalities.
GH is principally involved in the regulation of somatic growth, including cardiac development and function, and exerts its effects either directly or indirectly, by stimulating the production of IGF-I, that mediates GH action on peripheral tissues.
The relationship between GH/IGF-I and the cardiovascular system has been demonstrated by numerous experimental studies. In the heart, GH and IGF-I receptors are expressed in cardiac myocytes1 IGF-I causes hypertrophy of cultured rat cardiomyocytes2 and delays cardiomyocyte apoptosis.3 GH and IGF-I also have direct effects on myocardial contractility, by increasing the intracellular calcium content and enhances the calcium sensitivity of myofilaments in cardiomyocytes.4,5 Besides, endothelial cells have high-affinity binding sites for IGF-I6 and IGF-I stimulates nitric oxide (NO) production.7 Moreover, IGF-I, and its (IGFBPs participate to the inflammation-linked angiogenesis and repair processes following ischaemic events.8 A detailed description of the molecular basis of GH and IGF-I action in the cardiovascular system lies beyond the scope of this review (see Refs 1, 6, 8, 9).
Alterations of the GH/IGF-I contribute in determining cardiovascular disease as suggested by clinical studies reporting increased risk for cardiovascular morbidity and mortality both in GH deficiency (GHD) and excess.6,9 Epidemiological studies in the general population have also shown that IGF-I levels in the lower normal range are associated with an increased risk of ischaemic heart disease10,11,12 (IHD) and stroke.13–15
This review focuses on clinical consequence of GH–IGF-I axis on cardiovascular system in the general population and in the patients with GH excess or deficiency. Based on a PubMed search 1208 papers are available for ‘GH and cardiovascular disease’ and 1602 are those for ‘IGF-I and cardiovascular disease’. Thus in this review, I tried to offer a balanced view of some of the main clinical studies dealing with this subjects, either showing positive association between GH/IGF-I and the cardiovascular system or not, as modulated by my personal experience of 15 years of study in the field. Several excellent studies were, thus, excluded from quotation only because they did not offer a different view of the quoted papers.
The GH–IGF-I axis and cardiovascular disease in the general population
In recent years an association between low IGF-I and risk to develop IHD and stroke has been consistently demonstrated.
The risk to develop IHD was reported to be 1·27–1·94 times higher in the patients with lower IGF-I than in those with higher IGF-I10–12. Additionally, a higher risk for IHD was associated with IGFBP-3 in the high quartile (2·16 times higher)10 and with IGFBP-1 in the lowest quintile (3·1 times higher).12 Importantly, both IGF-I and IGFBP-1 (alone or in combination) were not related to any risk of mortality because of non-IHD.12 An increased risk of ischaemic stroke (2·1 and 2·3 times higher, respectively) was also observed in patients with IGF-I and IGFBP-3 levels in the bottom quartiles13 and low IGF-I levels were associated with poor outcome (mainly death) after stoke, independently from other clinical covariates.14,15
Additionally, the development of atherosclerosis was suggested to be delayed in presence of high free IGF-I levels16 and intima-media thickness (IMT) of common carotid arteries resulted to be lower in patients with higher IGF-I levels.17
Altogether, these data suggest that IGF-I levels in the higher quartiles of normal range are protective for atherosclerosis-associated cardiovascular events such as IHD and stroke. It is also important to point out that IGF-I levels in the middle-upper normal range was reported to be associated with several other conditions beneficial for the cardiovascular system such as reduced blood pressure18 and vascular tone19 increased insulin sensitivity20,21 and reduced prevalence of diabetes mellitus22 (Table 1). These effects, together with several others at the endothelial level such as increased survival of endothelial and smooth muscle cells23,24 plaque stability25,26 and anti-inflammatory actions27 should be considered to understand the possible contribution of IGF-I in reducing the risk for IHD and stroke.
Table 1. Cardiovascular and metabolic correlates of increased IGF-I in the general population
Incidence of IHD
Juul et al. 2002; Vasan et al. 2003, Laughlin et al. 2004
Incidence of stroke
Denti et al. 2004; Johnsen et al. 2005, Bondanelli et al. 2006
Van der Belt et al. 2003; Colao et al. 2005
BP and vascular tone
Gillespie 1997; Galderisi 2002
EC and VSMC survival
Patel 2001; Anwar 2002; Jia 2003
Paolisso 1999; Clemmons 2000
Incidence of diabetes
GH deficiency (GHD)
Findings in untreated patients
Patients with hypopituitarism are known to have reduced life expectancy with a 2-fold higher risk of death for cardiovascular disease compared with healthy controls.28–30 GHD has been considered the underlying factor of the increased mortality since all the other hormones were replaced at the time of death. A study from UK questioned the role of GHD on cardiovascular mortality, but only a small part of the entire cohort of the patients was studied for GH secretion.30 Although many other factors, such as excessive glucocorticoids or T4 replacement, gonadal steroids under-replacement, can potentially contribute to the increased cardiovascular mortality in hypopituitary patients, the direct effect of GHD is conceivable (Fig. 1). The GHD-mediated negative effects on the cardiovascular function are played both directly on the heart and endothelium, as previously mentioned, and indirectly via hypercoagulability, abdominal obesity, insulin resistance, high total and LDL-cholesterol and low HDL-cholesterol, atherosclerosis, decreased exercise performance, pulmonary capacity and endothelial function.31,32 Among all these indirect cardiovascular risk factors abdominal obesity, assessed simply by waist circumference or waist : hip ratio33 is a well-known negative predictor of subsequent coronary artery disease.34 Patients with GHD were consistently proven to be affected with increased centrally distributed adiposity and, additionally, with dyslipidaemia31,32 whose treatment has become crucial in primary and secondary prevention of cardiovascular disease.
The severity and duration of GHD was reported to correlate with adverse lipid35 and lipoprotein profile as there is an inverse association between IGF-1 and LDL-cholesterol.36 Additionally, early reports suggested that GHD patients could have a slight increase in blood pressure.37 An observational study conducted in Spain and recruiting almost 1000 GHD patients reported an increased prevalence of hypertension compared to the general population (22%vs. 15%).38 An increased sympathetic nervous system activity was found to be associated with a diastolic BP around 10 mmHg higher in GHD patients than in controls.39 Conversely, we did not find any increased prevalence of hypertension in 56 patients classified to have severe GHD40 compared to sex- and age-matched healthy controls. We found rather a decreased systolic blood pressure at peak physical exercise (Fig. 2).
Besides the cardiovascular risk factors mentioned above, patients with GHD were shown to have increased blood vessel IMT that is well known to represent one of the earliest morphological changes in the arterial wall in the process of atherogenesis. Markussis et al.41 first showed increased carotid IMT and atheromatous plaque prevalence in the common carotid artery in asymptomatic GHD patients. This result was later on confirmed in other series of GHD patients.42–44 Less distensibility of aorta45 and endothelial dysfunction46 were also demonstrated together with other reports on higher levels of fibrinogen, plasminogen activator inhibitor activity (PAI-1) and tissue plasminogen activator antigen (tPA) in GHD patients compared to controls.31,32
Finally, GHD modifies cardiac size and function. By echocardiography, a significant reduction in the thickness of the left ventricular (LV) posterior wall and of the interventricular septum, which resulted in a decrease of LV mass index, together with a decrease of the LV internal diameter, was reported in children, adolescents and childhood-onset adults with GHD.47–51 Other studies in adult patients developing GHD in adult age showed, however, similar cardiac function and morphology as and controls.52–54 Echocardiography is a method, however, not sensitive enough to reveal subtle deficiency in LV performance. We first analysed LV performance in response at peak exercise using the equilibrium radionuclide angiography. A remarkable impairment of LV performance was found in most adult patients with GHD, regardless their age and age of disease onset.55 More recently, we also reported that cardiac performance is correlated with the GH status; in fact, a significant impairment we found in patients with severe and partial GHD but not in non-GHD hypopituitary patients.40
In summary, besides well known abnormal body composition and dyslipidaemia, patients with untreated GHD have generally a reduced cardiac performance on peak exercise and increased IMT at major vessels. If GHD appears at a young age, patients can also have a reduced cardiac mass. Less consistent are data on increased blood pressure.
Findings after GH replacement
GH replacement beneficially changes body composition producing increase in lean body mass by 2–5 kg and decrease in body fat mass by 30% (approximately 4–6 kg of visceral fat).56 Improvement in body composition may be the single most important factor in reducing vascular risk, as estimated by the Framingham model to represent a 3–4% decrease in the incidence of coronary heart disease over 10 years.57 It has, however, been suggested that the majority of excess vascular risk associated with hypopituitarism is due to dyslipidaemia.58 A meta-analysis of blinded, randomized, placebo-controlled trials with low doses and long-duration GH treatment showed that GH replacement has beneficial effects on lean and fat body mass, total and LDL cholesterol levels, and diastolic blood pressure.59
In a few GHD adult patients GH replacement for 6–24 months was associated with reduced IMT at common carotid arteries up to levels recorded in controls in some studies.42,43,60,61 Withdrawal from GH replacement for 6 months in severe GHD adults, but not in adolescents62 was associated with an increase in IMT at the same vascular level and impairment of cardiovascular risk.61 Moreover, GH replacement induces a significant increase in flow mediated dilation, mediators of endothelial function and arterial compliance32 and of cardiac mass. Amato et al.47 first showed that GH replacement induced a 26% increase in the LV mass index and a 12% increase in the LV ejection fraction, that disappeared 6 months after GH discontinuation. It should be noted that the dose of GH employed during the last decade was reduced from 20 to 26 µg/kg body weight in the initial studies to 4–6 µg/kg body weight of modern ones. Increase of LV mass is, anyway, a rather common finding in the early GH replacement in adults63–67 as well as in children.50,51,68 Ter Maaten et al.69 noted that the hypertrophic effect of GH replacement during the first year of GH replacement was short-lasting and LV mass returned to normal 2 years after therapy continuation being similar to pretreatment values after 10-year replacement. A few studies did not report any significant change in cardiac mass and performance54,70–73 (Table 2). In a cohort of young GHD patients, we observed a significant increase of the LV ejection fraction at peak exercise after 12 month of GH replacement even if exercise-induced changes of LV ejection fraction remained significantly lower than controls after treatment.74 In comparison with a group of GHD patients who refused to undergo GH replacement, we confirmed that only the patients who had received GH replacement showed an improvement of cardiovascular risk parameters, cardiac mass and function parameters after 12 months.77 It should be emphasized, however, that GH replacement for 12 months was unable to completely normalize cardiac performance thus indicating that cardiac performance should be monitored in long-term studies. In fact, Chrisoulidou et al.75 reported a decrease of diastolic blood pressure and an improvement of diastolic filling persisting seven years after GH replacement. Improvement of systolic performance was not consistently reported: to be noted that echocardiography is not a method sensitive enough to reveal minimal changes in ejection fraction while other techniques like equilibrium radionuclide angiography, as we used, is more efficient because of exercise testing. Anyway, using echocardiography Ezzat et al.76 found data similar to those we reported in GHD adolescents.49
Table 2. Cardiovascular effects of GH replacement in GHD patients
In summary, GH replacement increases cardiac size, not exceeding normal values, improves cardiac performance, more evidently on peak exercise, and reduces IMT at common carotid arteries as documented by the majority of the studies. To weight the beneficial effect of GH replacement in terms of reduced mortality is a very important clinical question. The results of a large registry database suggest, however, that GH replacement therapy may be important in adults with GHD not only to improve general health and well-being but also to reduce the risk of premature mortality.78 As the individual changes in the cardiovascular risk factors in all the studies are small, the global benefit of GH replacement on cardiovascular mortality remains to be determined.
GH excess (acromegaly)
Findings in untreated patients
The most common feature of the acromegalic cardiomyopathy is concentric biventricular hypertrophy.6 Cardiac walls are generally thickened, but cardiac chambers are rarely enlarged due to relative increase of cardiac myocytes width.6 Ageing 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.6,79 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 year) to elderly (> 60 year) patients.6 Accordingly, the prevalence of left ventricular hypertrophy was higher in patients aged > 50 year (74·3%) than in younger ones (57% in patients aged 31–50 year and 35% in those aged < 30 year).6 Left ventricular hypertrophy is not negligible also in young patients with a presumed short duration of acromegaly.80–82 This suggests that cardiac hypertrophy is an early event in acromegaly, which worsens proportionately with the duration of disease activity. Arterial hypertension is likely the most important factor aggravating cardiac hypertrophy: the results of a multistep regression analysis showed that diastolic blood pressure was the best predictive factor of cardiac hypertrophy.83 It should be mentioned that patients with hypertension and diabetes are older than those with uncomplicated acromegaly. Since ageing in non acromegalic subjects is characterized by a slight increase in left ventricular hypertrophy84 it is conceivable that in acromegaly this phenomenon is emphasized. Short disease duration is associated with high heart rate and increased systolic output, altogether configuring the hyperkinetic syndrome.6 The acromegalic cardiomyopathy then progresses in overt cardiac hypertrophy with signs of diastolic dysfunction and/or insufficient systolic performance. This may then rarely develop into systolic dysfunction at rest and overt heart failure with signs of dilative cardiomyopathy in patients with untreated disease.6
Rhythm disturbances, such as ectopic beats, paroxysmal atrial fibrillation, paroxysmal supraventricular tachycardia, sick sinus syndrome, ventricular tachycardia and bundle branch blocks, are also more frequently recorded than in controls mainly during physical exercise.85,86 Cardiac valve disease is also underestimated: Lie and Grossman79 found mitral and aortic abnormalities in 19% of their autopsy series. We recently demonstrated a high prevalence of both mitral and aortic valve dysfunction in patients with active acromegaly:87 cardiac valve abnormalities were associated with left ventricular hypertrophy in our series and in the series published by Pereira et al.88 Hypertension is the major cardiovascular complication in acromegaly and is reported to affect approximately one third of patients. In a survey study of 200 patients with acromegaly studied at diagnosis, we found hypertension in 46% of patients significantly higher than in controls (25%), without any difference between men and women.89 The coexistence of glucose abnormalities and insulin resistance increase the probability to have coexistence of hypertension, which can also be due primarily to increased wall : lumen ratio in some vascular districts. Increase of the carotid IMT was observed in active as well as in cured patients with acromegaly but the prevalence of well-defined atherosclerotic plaques was not higher than in control subjects.90 Only mild increase of carotid IMT was conversely reported by Kasayama et al.91 however, since plasma IGF-I concentration was significantly higher and the prevalence of hypertension was significantly lower in the 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 patients. We could not find, however, any correlation between IGF-I levels and carotid IMT.90–92 Laser Doppler flowmetry has also confirmed endothelial dysfunction.93,94 The existence of other negative factors, such as glucose intolerance, dyslipidaemia, smoking habitus, further impair vascular relaxation.
In summary, patients with acromegaly are generally diagnosed when cardiac hypertrophy has already developed. According with the duration of untreated GH and IGF-I excess other cardiac signs can be revealed on echocardiography such as cardiac valve calcification, fibrosis and regurgitation and diastolic dysfunction. Arrhythmias are generally of minor clinical relevance, but they should be recognized and treated as considered a cause of sudden death.6 Less consistent data have been provided on increased IMT at major vessels. Cardiac performance on peak exercise has been demonstrated to be impaired, mostly in patients with longer disease duration.6
Findings after GH suppression
Control of GH and IGF-I excess can arrest the progression of cardiac disorders in line with epidemiological data showing that disease control is associated with a consistent reduction of cardiovascular mortality.95 Surgical cure by adenomectomy was reported to reduce cardiac mass and improve diastolic filling in small patients series.96–101 Only one study has reported the results of radiotherapy: this rather old study reported an impairment of cardiovascular profile but no data on aggravation of post-radiotherapy hypopituitarism were available.102 The treatment with somatostatin analogues has widely been reported to successfully improve cardiovascular parameters in a number of studies. Reduction of cardiac mass and improvement of diastolic filling occur in patients treated with s.c. octreotide, lanreotide and octreotide-LAR.103–116 Diastolic and systolic improvement is more evident in patients achieving disease control, while those not controlled by therapy had no response98,112,113 or, even, further impairment of their cardiac function.98 The age of the patients also determines the positive response to the treatment. In 22 patients successfully controlled for 1 year by octreotide-LAR, we observed the disappearance of left ventricular hypertrophy in 100% of patients aged below 40 years and only in 50% of those aged above 40 years.115 In addition, the left ventricular ejection fraction response at peak exercise significantly increased only in younger patients, being restored in 80% of young and in 50% of middle-aged patients.115 These observations suggest that the acromegalic cardiomyopathy is more likely to be reversed in younger patients with short disease duration, whose disease activity is successfully controlled by 12-month treatment with octreotide-LAR. At further support of this concept, patients aged below 40 years and with disease duration shorter then 5 years had reversal of cardiac hypertrophy, impairment of ejection fraction at peak exercise as well as decrease of IMT at common carotid arteries.82 This is also in line with the previously mentioned negative effect of prolonged disease duration on the severity of the cardiomyopathy: thus both the diagnosis and the treatment beginning should be as early as possible in order to preserve left ventricular performance. Recent data reported by Pivonello et al.117 using the GH receptor antagonist pegvisomant, have reinforced the importance of IGF-I control to improve cardiac mass and performance in patients with active acromegaly. The results of a beneficial effect of somatostatin analogues on several aspects of the acromegalic cardiomyopathy have been confirmed by a very recent meta-analysis.118 Maison et al.118 documented that somatostatin analogue treatment was associated with significant reductions in the heart rate, the left ventricular mass index, and the ratio of the E-wave and A-wave peak velocities of the mitral flow profile, also associated with improved exercise tolerance and trends toward beneficial effects on left ventricular ejection fraction. More recently, decrease in myocardial fibrosis has been reported after treatment of acromegaly.119,120
Most study dealing with the effects of somatostatin analogues treatment reported decrease in heart rate (Table 3), an effect of relevance when the drug was given preoperatively.121 Somatostatin analogues were also proven to reduce the QT interval, considered an established risk factor for potentially fatal cardiac arrhythmias122 and to improve the arrhythmic profile.116 This effect can be due not only to the decrease of GH and IGF-I levels but also to a direct effect of octreotide, that was shown to have direct effects on the conduction system.123 Besides this effect on cardiac rhythm, somatostatin analogues have recently been suggested to play direct effect on cardiac performance. In fact, Smith et al.124 have shown that somatostatin receptors type-1 (sst-1), sst-2, sst-4 and sst-5 were coexpressed in both atrial and ventricular tissue. Human cardiac myocytes expressed mRNA for only sst-1 and sst-2, while human cardiac fibroblasts expressed all four subtypes. These findings raise the possibility of a direct effect of somatostatin analogues on the heart that should be more carefully investigated.
Table 3. Cardiovascular effects of treatment of acromegaly
Year of publication
RT, radiotherapy; Oct, subcutaneous octreotide; LAN, lanreotide slow release; LAR, Octreotide-LAR; SSA, somatostatin analogues (this label indicates the use of different analogues in the treatment of acromegaly); P, patients; ↑, increase; ↓, decreased; ↔, unchanged.
No improvement of cardiac valve disease after treatment of acromegaly has been reported by van der Klaauw et al.125 and no consistent data are available on hypertension and endothelial dysfunction. In fact, no change of blood pressure was revealed in the meta-analysis of the literature.118 It is possible, however, that only hypertensive patients should be analysed to evaluate potential beneficial effects of somatostatin analogues on blood pressure. Finally, some reduction of IMT at common carotid was reported in a few patients after lanreotide treatment.92
Table 3 summarizes the results of different treatments of acromegaly on cardiovascular parameters.
In summary, control of acromegaly enables the recovery from cardiac hypertrophy and improves diastolic dysfunction and systolic dysfunction at peak exercise, mostly in young patients who had had a presumably short exposure to GH and IGF-I excess. In contrast, cardiac valve disease does not seem to be substantially changed by controlling hormone hypersecretion, but data on this issue are still scant. It is clinically relevant that cardiomyopathy can led eventually to diastolic heart failure, sometimes very hard to treat and control, if acromegaly is left untreated. In 10 patients with acromegaly and chronic congestive heart failure (out of 330 consecutive patients treated in two French and Belgian centres since 1985), Bihan et al. have even questioned the impact of effective suppression of GH and IGF-I on long-term survival in these patients.126 Since these are rare patients (< 3%) the final effect of tight GH and IGF-I control in patients with congestive heart failure need to be re-evaluated.
Abnormalities of the GH–IGF-I axis in adults are associated with a number of detrimental factors for the cardiovascular system. Reductions in the level of GH and/or IGF-I is associated with alterations of body composition, lipid profile and coagulation pattern leading to increased prevalence of atherosclerosis, IHD and stroke. GH replacement in patients with GH deficiency has been substantially proven to reduce visceral fat, that consequently improves insulin sensitivity, and to reduce LDL-cholesterol and triglycerides levels, so improving the pro-atherosclerotic profile. Besides these effects, decrease of the IMT at major arteries, improvement of endothelial dysfunction, considered surrogate markers of atherosclerosis, and improvement of cardiac performance have been reported. These beneficial effects of GH replacement in adult GHD need to be confirmed in terms of life expectancy of these patients. GH excess in acromegaly, on the other hand causes a specific cardiomyopathy, considered one of the most severe complications of the disease. Control of GH and IGF-I excess is reportedly associated with improvement of left ventricular hypertrophy, diastolic dysfunction and arrhythmias. Other aspects, such as hypertension, cardiac valve disease and endothelial dysfunction have been studied in less detail and need to be re-evaluated. Anyhow, improvement of cardiomyopathy is observed mainly in young patients with short estimated disease duration, suggesting that early diagnosis is made and prompt treatment of acromegaly is performed.
The author thank the scientific committee of the British Endocrine Society for awarding with the European Gold Medal lecture whose content is reported in this review. The author wishes to thank Professor John Wass (University of Oxford, UK) for his valuable help in revising this review.