To view article on Medscape click here
From Endocrine Practice
Cardiovascular Risk in Patients with Growth Hormone Deficiency: Effects of Growth Hormone Substitution
Albert G. Burger, MD;1 John P. Monson, MD;2 Anna M. Colao, MD;3 Anne Klibanski, MD4
Abstract and Introduction
Objective: To review the literature on the increased cardiovascular risk in patients with growth hormone (GH) deficiency and the positive effects of GH replacement.
Methods: We analyze the factors that contribute to cardiovascular risk in GH deficiency, including body composition and lipid profile, and summarize GH treatment strategies and results described in the literature.
Results: The prominent clinical finding in patients with GH deficiency is the increased abdominal fat, even in patients with normal weight. Cardiac ejection volume tends to be decreased, and arterial distensibility is diminished. The lipid status is also worsened, accompanied by increased inflammatory markers, such as highly sensitive C-reactive protein. Typically, GH treatment reduces visceral fat and increases muscle mass, changes that diminish cardiovascular risk. Because of direct effects as well as increased hemodynamic performance and increased blood volume, cardiac performance is improved. With GH therapy, total cholesterol and low-density lipoprotein levels decrease by 10% to 20%, and inflammatory markers such as C-reactive protein decline. Carbohydrate metabolism during moderate to long-term treatment is minimally affected, although obese patients with GH deficiency on rare occasion may have hyperglycemia or even diabetes.
Conclusion: The relevance of the beneficial effects of GH on the cardiovascular system is strongly suggested but not fully proved. The results in a large cohort of GH-treated patients (the KIMS or Pharmacia and Upjohn International Metabolic Surveillance database) demonstrated no difference in cardiovascular risk in comparison with that in a control population after a mean of 3 years of treatment.
Growth hormone (GH) deficiency in adults is frequently attributable to tumors affecting the hypothalamic-pituitary region. Deficiency of GH, however, is not only seen in this clinical context. GH is among the most sensitive pituitary hormones to irradiation, traumatic brain injury, and possibly chemotherapy. Clinically, GH deficiency may easily be missed. In a 5-year follow-up study of patients having received classic external radiotherapy for pituitary or other brain lesions, GH deficiency was noted in almost 100% of the cases. Mainly gonadal function and, somewhat less frequently, thyroidal and adrenal function were also affected. Investigators have established that patients with panhypopituitarism have an increased cardiovascular (CV) risk. Conventional therapy with thyroid hormones, cortisol, and sex hormones? particularly gonadal substitution in women?is certainly effective in decreasing the CV risks. Nonetheless, it has been recognized that these patients, even though treated with thyroid and steroid hormones, continue to have a higher than normal CV risk, the risk in women exceeding that in men.[2,4-6] In addition, cerebrovascular disease was more frequent, particularly in women. In 1990, Rosen and Bengtsson first reported that, despite classic substitution therapy (excluding GH replacement), patients with panhypopituitarism remained at increased CV risk and suggested that GH deficiency might have an etiologic role. Other investigators, however, have suggested that external irradiation or the underlying pituitary pathologic condition may have an etiologic role in cerebrovascular disease.
Lean Body Mass Versus Adipose Tissue
The increased CV risk in overweight patients is particularly dependent on a high proportion of abdominal (visceral) fat in comparison with total body fat.[8-12] Insulin resistance (IR), increased inflammatory markers, and microvascular and macrovascular disease are common complications. Patients with GH deficiency demonstrate a high prevalence of obesity and share most of the risk factors of those with simple obesity. Blood pressure, however, is not necessarily increased in patients with GH deficiency.
Nonobese and even lean patients with GH deficiency have a high ratio of abdominal to total body fat.[5,11] They are at a higher CV risk than an age- and weight-matched population. It is recognized that these subjects have a lean body mass deficit of approximately 3 to 4 kg and an increase of 1 to 2 kg in abdominal fat. This pattern is frequently observed in women with panhypopituitarism but is also seen in men.
Vascular Motility and Plaque Formation
The poor CV profile of patients with hypopituitarism and GH deficiency was further substantiated by observations of vascular mortality, plaque formation, increased intima-media thickness (IMT), decreased production of nitric oxide (NO), abnormal lipid profiles, inflammatory markers, and development of IR. The macrovascular abnormalities were first investigated by ultrasonography in 1992 and 1993.[13,14] These early studies documented that atheromatosis of the abdominal aorta and the femoral and carotid arteries was more pronounced in patients with GH deficiency than in the control population. In addition, the distensibility of these arteries, measured by ultrasonography, was appreciably reduced. Several studies have confirmed and extended these findings.[15-21] Apparently, there was no sex-related effect on the prevalence or severity of atherosclerosis in adult patients with GH deficiency. The distensibility of the forearm arteries has been studied by using an ischemic-hyperemic stress or by applying pharmacologic dilatation of the vessels. All studies confirmed an increased stiffness of the arteries in comparison with that in the control population.
The pathophysiologic basis for these findings has been documented by the remarkable work of Boger et al. They were able to demonstrate that GH was responsible for endothelial NO production. NO is not only a potent vasodilator but also an inhibitor of low-density lipoprotein (LDL) oxidation in the arterial wall, superoxide radical elaboration, smooth muscle proliferation, and platelet aggregation and therefore has a potent antiatheromatous action. Most likely, many if not all of these effects are mediated by insulin-like growth factor-I (IGF-I).
GH and IGF-I have a direct positive effect on cardiac morphologic findings and function, which are essential for cardiac performance. In experimental models, hypophysectomy induces a decrease in size of several organs, including the heart, which is reversed by administration of GH. In human GH deficiency, this relationship is well documented in childhood-onset disease; in such patients, reduced systolic performance is responsible for the hypo-kinetic syndrome detected in some of these children. In adult-onset GH deficiency, some studies have also described a decrease of left ventricular (LV) muscle mass and cardiac ejection volume[25-28] whereas others have not,[29,30] early adulthood-onset disease being more subject to these changes than onset in middle age or in elderly patients. GH treatment tends to improve cardiac function. This result has been documented by studies in childhood-onset and adult-onset GH deficiency[31,32] and confirmed by a large meta-analysis that included all the published randomized controlled trials on cardiac aspects of GH replacement. That article reports a significant positive effect on LV mass, interventricular septum, LV posterior wall, LV end-diastolic diameter, and stroke volume after GH therapy in adults.
Lipids and Lipoproteins
Meta-analysis of the lipid status of patients with GH deficiency clearly indicates the presence of an adverse lipid profile. Among elderly patients, a substantial proportion have elevated total cholesterol and LDL levels, and high-density lipoprotein cholesterol concentrations tend to be decreased. In some studies, the atherogenic apolipoprotein (apo) B-100 was evaluated and tended to be increased.[3,33,34] Triglyceride levels also tend to be increased.[5,10] This lipid profile is consistent with an increased risk of atherosclerosis and CV events.
Atherosclerosis is an inflammatory process, and markers such as highly sensitive C-reactive protein (hsCRP) or interleukin-6 (IL-6) are highly sensitive indicators if other sources of inflammation, such as infections, are excluded.[9,35-37] Examples of other less frequently used markers are serum myeloperoxidase, fibrinogen, sialic acid levels, cystatin C, and adhesion of monocytes. In patients with GH deficiency, hsCRP values and myeloperoxidase are considerably increased, whereas fibrinogen has only a tendency to be increased.
Carbohydrate and Amino Acid Metabolism
Despite normal blood glucose levels, patients with GH deficiency tend to have an increased serum insulin concentration and evidence of IR, consistent with the augmented abdominal (visceral) adiposity. IR has been documented by using various techniques, including glucose clamp studies.[8,40,41] Interestingly, patients with acromegaly also show a tendency for having IR and diabetes mellitus, despite a reduction in abdominal adiposity; this finding suggests that both an excess and a lack of GH result in IR.
Elevated serum homocysteine levels are considered an independent risk factor for increased CV events. In placebo-controlled studies in men, GH treatment was found to decrease fasting homocysteine levels. These changes did not correlate with changes in C-reactive protein (CRP), IL-6, or insulin levels.[42-44]
In summary, patients with hypopituitarism, despite adequate treatment of pituitary deficiencies other than GH deficiency, have the morphologic features and many of the biochemical findings that characterize patients with a high CV risk. Investigators have estimated that there are approximately a 2-fold increase in CV risk in men beyond the age of 50 years and a 3.5 to 4-fold increase in women of that age-group. Importantly, the increased CV risk also seems to be relevant for patients with GH deficiency who have a normal body mass index. The most important factors in GH deficiency and their changes with GH treatment are summarized in Table 1 .
Biochemical Diagnosis of GH Deficiency in Adults
The definition of partial degrees of GH deficiency remains controversial, although the diagnosis is rarely difficult or contentious in patients with structural pituitary disease. Despite the availability of highly sensitive assays that measure GH down to 0.02 ng/mL, the diagnosis cannot be based on basal GH values. Among the functional tests, the insulin tolerance test remains the "gold standard." It is commonly agreed that, after profound hypoglycemia (plasma glucose levels <2.2 mmol/L or 40 mg/dL), severe GH deficiency is confirmed if serum GH values do not exceed 3 ng/mL.[45,46] Criteria for patients with incomplete GH deficiency do exist but are beyond the scope of this review. The test has obvious limitations in elderly patients, in patients with diabetes mellitus, and in obese patients, and it is contraindicated in patients with epilepsy or ischemic heart disease. Other validated tests are the glucagon and the growth hormone-releasing hormone/arginine test. Another well-tolerated test, the hexarelin test, which works through the ghrelin receptor, is not universally available.
Most specialists would propose an IGF-I value in the range of the mean value + 2 SD as a target for adequate treatment, taking into account the age-adjusted reference range. GH treatment should be initiated with use of a low dose (to prevent side effects) and then gradually increased. Maintenance doses differ for childhood-onset or adult-onset disease as well as age and sex of the patient. In men, dosages range from 0.3 to 0.6 mg daily; in premenopausal women or those receiving orally administered estrogens, a higher dosage?from 0.4 to 1 mg daily[49,50] ?is usually required. The dose cannot be strictly calculated on the basis of body weight or surface area but must be altered in accordance with the age- and sex-adjusted normalized serum IGF-I values. Thus, adjustment of treatment is necessary after onset of the menopause or after discontinuation of estrogen therapy. For achieving a full effect on many variables such as body composition, maintenance treatment must be continued for at least 6 months. For other end points such as bone density, a period of 18 months may be needed to demonstrate a significant effect, and improvement may continue beyond 18 months of treatment. Of note, the early trials of GH replacement used higher doses, which occasionally had to be reduced during the experimental protocol because of side effects. This factor can make the interpretation of results more difficult.
Contraindications to GH treatment are active malignant disease, benign intracranial hypertension, and diabetic retinopathy. Patients with a past history of a cured malignant condition, other than skin cancer, should undergo thorough evaluation for the risks versus benefits of treatment. Concerns about GH treatment and malignant lesions stem from the positive association observed between IGF-I levels and risk of cancer, particularly prostate cancer, in the normal population. Men with the highest quartile of IGF-I values had a relative risk of prostate cancer of 4.3. Despite these observations in normal populations, surveillance studies have thus far not been able to show any increased risk of malignant involvement in treated patients with GH deficiency in comparison with those not treated with GH.
Effects of GH Treatment
Clinically, the most striking change in CV risk factors with GH treatment is the improvement of body composition, a 1- to 3-kg decrease of abdominal or visceral fat and an increase in muscle mass, together with improved physical performance (Fig. 1).[10,12,52-55] These effects are also seen in elderly patients. The increase of cardiac mass?in particular, LV mass?was noted in several studies, including in adult patients with childhood-onset GH deficiency. The increase of LV mass was within physiologic limits; pathologic cardiac hypertrophy attributable to GH substitution therapy has not been reported. After discontinuation of GH treatment, the positive effect does not persist, and within 6 months, LV mass returns to pretreatment values.[30,32,57-60]
Figure 1. (click image to zoom)
Effects of growth hormone (GH) treatment in patients with GH deficiency. + = favorable effects of GH; ? = slightly positive or neutral effects.
The effect on cardiac mass is associated with enhanced cardiac performance. Improvement of both systolic and diastolic filling function has been observed. In addition to the changes in cardiac mass, hemodynamic improvements due to increased plasma volume and red blood cell mass also contribute to better cardiac performance.[22,25,61] The improvement of cardiac performance could be documented at rest, at peak exercise testing, and during sustained exercise. Confirmation of these observations was obtained in a large meta-analysis of all the available published studies on cardiac aspects of GH replacement. After therapy with GH in adult patients, significant positive effects were noted on LV mass, interventricular septum, LV posterior wall, LV end-diastolic diameter, and stroke volume. These results apply obviously to physiologic GH substitution because an excess of GH, as seen in patients with acromegaly, may induce pathologic cardiac hypertrophy, predisposing to arrhythmias and ultimately cardiac failure.
An improvement in the lipid profile is often seen after GH replacement therapy. A meta-analysis of several studies documented the effects of GH therapy on total cholesterol: the changes are particularly prominent in elderly patients with increased serum cholesterol levels and are not clearly seen in young subjects with normal cholesterol levels. Apo B-100, the crucial protein for cholesterol metabolism and a known independent risk factor for CV disease, has been shown to decrease after GH therapy. LDL cholesterol concentrations decrease, although with the currently used lower substitution doses of GH the effect is more modest and not seen uniformly in randomized placebo-controlled trials. In comparison with statins, the effect of GH may be less impressive because the mean decrease in total cholesterol level is about 10% to 20% with GH. It is, however, not negligible because it is an additive effect. The effects on serum triglycerides have been variable, some authors reporting a decrease and others not finding this result.[5,10]
Arterial distensibility and plaque formation are also favorably affected by GH treatment. These findings have been documented by echo Doppler ultrasound investigations of the carotid arteries. IMT has also been measured. The IMT decreased significantly (approximately 0.1 to 0.2 mm) with GH treatment. Nonetheless, the effect was dependent on continuous treatment; stopping GH treatment for 6 to 12 months resulted in a recurrent increase of IMT. GH treatment is also associated with improved peripheral vasodilatation. This outcome can be documented by testing the flow-mediated vasodilatation of a brachial artery (see previous material). These results point to endothelium-derived NO and its crucial role in vasodilatation. Production of NO, judged from urinary nitrate and cyclic guanosine monophosphate production, shows a sustained improvement with GH treatment. In light of these effects, it is not surprising that systolic and diastolic blood pressure measurements decrease slightly but significantly in hypertensive and normotensive patients in response to GH replacement therapy.
The role of inflammation is critical in the pathogenesis of atheromatosis. CRP, by interacting with endothelial receptors, accelerates vascular inflammation and the development of atheromas. Its effect is mediated mainly through IL-6, a proinflammatory cytokine. hsCRP values are good indicators of CV risk; values of less than 1 mg/L, 1 to 3 mg/L, or greater than 3 mg/L indicate low, average, and high risk, respectively. Most patients with GH deficiency have CRP values in the range of moderate to high risk. By controlling for other risk factors, it can be estimated that the risk for an acute CV event is increased approximately 4-fold. Most, but not all, published studies demonstrate a significant decrease of CRP values with GH replacement therapy. IL-6, myeloperoxidase, and fibrinogen are also inflammatory markers of vascular risk, although they have not been studied as extensively as CRP. It is noteworthy that only a few drugs have been shown to decrease CRP values, the best known being the statins. GH is also a cytokine, and its receptor belongs to the family of the cytokine receptors. Intracellular activation occurs through STAT4 (the signal-transducing activator of transcription protein 4), a well-known pathway for cytokines. It is tempting to speculate that by interfering with the action of proinflammatory cytokines, GH reduces or even reverses IMT and plaque formation.
Although one might be impressed by the many favorable effects of GH substitution in patients with GH deficiency, it must be remembered that acromegaly is associated with an increased incidence of cardiac insufficiency and arrhythmias, development of IR, and overt type 2 diabetes. Interestingly, patients with severe GH deficiency also show signs of IR. During early GH treatment, glucose tolerance may even worsen before improving. Typically, it normalizes after several months of treatment, probably a consequence of decreasing abdominal adiposity. With long-term GH treatment in nonobese patients, serum insulin, glucose levels, and the homeostasis model assessment index (insulin/glucose ratio as an index of IR) tend to normalize. In obese patients with GH deficiency, however, an increase in blood glucose levels, in rare cases within the diabetic range, has been observed. Thus, carbohydrate metabolism during long-term GH therapy should be monitored.
For clinicians, it is becoming clear that the substantially increased CV risk in patients with hypopituitarism and GH deficiency can be ameliorated by adequate treatment with GH. The effects of GH replacement have been carefully monitored in the KIMS (Pharmacia and Upjohn International Metabolic Surveillance) database, which includes approximately 10,000 patients. This database did not specifically address the question of CV risk; thus, the possibility exists of inadequate and potentially biased patient recruitment. In these GH-treated patients, the incidence of CV events did not differ from that in the control population after a mean of 3 years, a finding that suggests that GH replacement may be antiatherogenic. There was also no evidence of increased incidence of malignant disease.[66-69] For this reason, in many countries, GH substitution therapy is routinely accepted and reimbursed by medical insurance companies. Continued monitoring of GH-treated patients should also facilitate accumulation of evidence of the CV benefit of this treatment.
In summary, there is no doubt that adult patients with hypopituitarism and GH deficiency have an increased CV risk (odds ratio of 2 for men and 3.5 to 4 for women). The risk is not only limited to overweight patients but also seen in those with normal body mass index. In the follow-up of these patients with GH treatment, body weight and abdominal adiposity, readily documented by recording the waist-to-hip ratio, are certainly useful in daily clinical practice. Among the biochemical factors, hsCRP, total cholesterol, LDL, and apo B levels can be monitored. Moreover, in patients with known CV disease, IMT may be a helpful marker. The effectiveness of GH treatment must be monitored carefully by measuring the age- and sex-adjusted IGF-I values, and the glucose status must be followed throughout the entire treatment period.
Improvement of CV function should not be expected before at least 6 months of stable treatment.