Cardiovascular disease and risk factors in adults with hypopituitarism

In 1992, Cuneo et al. (1992) reviewed, in this journal, the consequences of growth (GH) deficiency during adult life and coined the term ‘Growth hormone deficiency (GHD) syndrome in adults’. Limited information was available then on cardiovascular disease and risk factors in adults with hypopituitarism. More data have been published since on cardiovascular morbidity in symptom-free hypopituitary adults. Several groups have also published results on the effects of GH therapy on cardiac structure and function and on the cardiovascular risk factors. We shall review here the epidemiological and experimental data on cardiovascular morbidity and mortality in adult hypopituitarism. We shall also examine the prevalence of cardiovascular risk factors in this group of patients and their potential causes focusing on the possible role of GH deficiency and the effects of GH replacement. Life expectancy in hypopituitary patients on routine replacement has been previously assumed to be normal (Hall, 1972). GH deficiency was proposed by others to confer a protective effect against risk of vascular disease from coexisting hyperlipidaemia (De-Gennes et al. 1979). Rosen & Bengtsson (1990) were the first to demonstrate that life expectancy was decreased (Table 1). They examined the records of 333 patients who attended the endocrine clinic in Gothenburg with a diagnosis of hypopituitarism between 1956 and 1987. 104 patients died during the study period and the mortality rate was significantly higher than that in the Swedish population (observed/expected; 1.8, P < 0.001). The excess mortality reflected deaths from vascular disease (Table 1). Myocardial infarction, ischaemic heart disease with congestive cardiac failure and cerebrovascular disease were the most frequent causes of vascular death. Another report from Lund, Sweden (Bulow et al., 1997) on 344 patients diagnosed between 1952 and 1992 demonstrated increased overall mortality (SMR 1.75; 95%CI, 1.4–2.19). It confirmed increased deaths, and in this study, this was particularly due to cerebrovascular disease (SMR 3.39; 95% CI, 1.41–1.88). The increase in cardiovascular mortality was greater in women than in men (SMR. 2.39 vs 1.54). Wüster et al. (1991) examined vascular disease in 122 patients with hypopituitarism in Heidelberg, Germany. The frequency of arteriosclerotic events (myocardial infarctions or strokes) was also higher than that anticipated for the German population. Using the Framingham model, Rosen et al. (1998) calculated the risk of coronary heart disease events from abnormalities of plasma lipids to be 130 events/1000 individuals in GHD men vs 40 events in male controls/1000 individuals and 47 events/1000 individuals in GHD women vs 10 events/1000 individuals in female controls. However, these calculations were though to have very major methodological flaws (Erfurth & Hagmar, 1998). Further attempts at calculating the risk using the frequency of the multiple risk factors observed in hypopituitarism would be welcome. Excess vascular mortality has not been universally observed. Bates et al. (1996) reviewed the case notes of all patients who were investigated for hypopituitarism at North Staffordshire Hospital, UK (Table 1). Of 172 patients who had at least one biochemically demonstrable hormone deficiency, 60 died during the study period. The overall mortality was increased (observed/expected; 1.73, P < 0.01). Vascular disease accounted for 42% of deaths but in this smaller study no significant excess vascular mortality was observed (observed/expected 1.35; P = 0.11). A preliminary report from Birmingham, UK on mortality in a cohort of 349 patients (200 men and 149 women) with pituitary disease diagnosed between 1968 and 1992 has been presented (Bates et al., 1997). The vascular deaths so far has been relatively low (53 men and 29 women died during the study period at a median age of 64 years). The overall observed/expected ratio was increased but there was no excess vascular mortality in either sex. Assessment of adult subjects with GH deficiency since childhood showed no evidence of increased vascular disease (Merimee et al., 1973; Merimee, 1978; Libber et al., 1990). Morbidity and mortality in two other series of patients who had received GH in the UK and Canada were mostly related to recurrence of the underlying pituitary disease or its metabolic complications (Buchanan et al., 1991; Taback et al., 1996). There is therefore evidence for reduced life expectancy in adult-onset hypopituitarism mainly due to vascular disease in some but not all studies. There are not enough data on vascular risk on those who develop GH deficiency during childhood and long-term follow-up studies of these patients into adult life are required. Excess GH secretion in acromegaly is associated with increased heart mass and this decreases following reduction of circulating GH levels (Hirsch et al., 1969; Martins et al., 1977; Mather et al., 1979; Hayward et al., 1987). In experimental animals, hypophysectomy causes a decrease in the size of several organs, including the heart (Hjalmarson et al., 1969) and this is reversed with GH administration (Beznak, 1954). Several groups have examined the cardiac structure and function in adults with GH deficiency. They have employed two-dimensional echocardiography, Doppler techniques, radionuclide angiography and exercise tolerance test protocols. The left ventricular posterior wall thickness (LVPWT), interventricular septal thickness (IVST), left ventricular internal diameter (LVID) and fractional shortening (FS%) were found to be reduced in the patients in two studies from Italy (Amato et al., 1993; Merola et al., 1993) (Table 2). In our own patients, LVPWT and IVST were lower in patients than in controls but left ventricular mass index (LVMI), LVID and ejection fraction (EF%) were not different (Table 2). In other 3 studies (Thuesen et al., 1994; Dunne et al., 1995; Valcavi et al., 1995) there was no significant difference between patients and controls in the echocardiographic measurements of cardiac structure and systolic function (Table 2). In 20 GH deficient patients studied by Nass et al. (1995) the measurements were within the normal range for their laboratory although matched controls were not studied. We have demonstrated a positive relationship between left ventricular mass, total-body fat-free mass and serum IGF-I concentrations in hypopituitarism (Shahi et al., 1992; Beshyah et al., 1996), suggesting that GH in man does influence cardiac size and components of body composition to a similar extent. Left ventricular diastolic function has been assessed in five studies (Caidhal et al., 1994; Cittadini et al., 1994; Thuesen et al., 1994; Valcavi et al., 1995; Beshyah et al., 1996). In our patients, prolonged isovolumic relaxation time (IVRT) and decreased early/atrial peak velocity ratio (E/A ratio) were observed in 8 of 39 patients and the patients differed statistically from controls (IVRT: 94.1 ± 22.9 vs 83.3 ± 15.7 ms, P < 0.02; E/A: 1.25 ± 0.38 vs 1.46 ± 0.47, P < 0.04, in patients and controls, respectively). Caidhal et al. (1994) reported that left ventricular filling was abnormal in 3 out of 10 patients but Thuesen et al. (1994) found no significant difference between patients and controls (E/A ratio: 1.9 + 0.5 vs 1.7 ± 0.3, NS). E wave deceleration time, another marker of left ventricular filling, was increased in 20 patients compared with 20 controls (224.5 ± 8.0 vs 181.6 ± 9.9 ms, P < 0.01) although the E/A ratio was not abnormal in this study (Valcavi et al., 1995). Using radionuclide angiography Cittadini et al. (1994) demonstrated reduced peak filling rate in hypopituitary patients, again indicating impaired left ventricular diastolic function. The clinical significance of abnormal diastolic function is not clearly understood. In other groups of patients abnormal diastolic function indicates early myocardial disease and/or ischaemic damage and may predate systolic abnormalities and heart failure (Clarkson et al., 1994). Cardiac performance during exercise has been investigated in a limited number of studies. Hypopituitary patients could exercise for a shorter time and tolerate lesser workload than controls (Beshyah et al., 1996; Longobardi et al., 1998). Cardiac function is a very important but not the only determinant of exercise capacity. There was a higher frequency of exercise-induced ST segment depression greater than 1 mV in our patients than in normal controls (74% vs 28%, P = 0.0001) (Beshyah et al., 1996). All those with abnormal exercise-induced ST segment depression were in the group with left ventricular diastolic dysfunction on echocardiography (Shahi et al., 1992; Beshyah et al., 1996). In asymptomatic subjects, such electrocardiographic changes during exercise are associated with an increased risk of developing ischaemic heart disease on long-term follow-up (McHenry et al., 1984). Abnormal left ventricular systolic function (cardiac index and EF%) during the peak of exercise has been reported in hypopituitary patients compared with the controls (Cittadini et al., 1994). In the hypopituitary patients, EF% demonstrated an abnormal response (decrease) during exercise (Cittadini et al., 1994; Longobardi et al., 1998) and the abnormality correlated with the duration of hypopituitarism (Cittadini et al., 1994). The cardiac dysfunction was present in both childhood and adulthood-onset hypopituitarism when the two groups were analysed separately (Longobardi et al., 1998) Our own group has investigated the carotid and femoral arteries of hypopituitary patients and matched controls by high-resolution ultrasonography (Markussis et al., 1992). Carotid intima-media thickness was greater in patients than controls (Table 3). The percentage of patients with one or more atheromatous plaques and the percentage of individual arteries with a plaque were both higher in the patient group (Table 3). Increased intima-media thickness has been also reproduced in another study of 14 childhood-onset GHD adults (Capaldo et al., 1997) (Table 3). Increased intima-media thickness is indicative of early atheroma in epidemiological studies (Salonen & Salonen, 1993) and abnormalities in carotid arteries disease in other particularly the coronary arteries et al., 1994). The index calculated from changes in diameter and the is significantly higher in hypopituitary patients than in controls (Markussis et al., 1997) (Table 3). have demonstrated that the by wave Doppler is significantly reduced (Table suggesting that the also is et al., 1994). using this is reduced in with an increased risk of vascular such et al., and et al., the other Dunne et al. (1995) reported that assessed by was not different in controls and hypopituitary patients on of increased on reduction of the to a effect the of A effect from the reduced was In this study, both systolic and diastolic and by a similar to controls in response to several cardiovascular suggesting that cardiovascular were et al., 1995). The of evidence available therefore that hypopituitarism is associated with abnormalities of the with early The on is limited at least in the early increased frequency of was observed in one series of hypopituitary patients in significantly greater than that in the population at et al., Rosen et al. also found to be increased in frequency in adult-onset hypopituitary men compared to the Swedish population vs P < although the in women not significance vs In an found no significant difference in the prevalence of in adults of childhood-onset GH deficiency with compared with normal controls vs NS). In other smaller series including matched controls, has been either similar in patients and controls (Markussis et al., 1992; et al., 1993) or lower in the patient group et al., 1993; Thuesen et al., 1994; Longobardi et al., 1998). In not significantly to 10 in adult patients on routine replacement following hypophysectomy et al., changes in and effects of replacement were studied by other The by increased significantly in hypopituitary patients = on either or replacement was associated with significant in et al., 1994). In another study, the during the and time systolic and to be lower in hypopituitary patients than in normal controls, though not significantly so et al., 1995). The of data therefore that is not a major of hypopituitarism. The increased prevalence compared with normal may the patient increased for The excess vascular disease not be to The of an risk for vascular disease has been although the of body is more is associated with increased vascular mortality and morbidity et al., et al., and risk factors for vascular disease such hyperlipidaemia and et al., et al., is in hypopituitary adults et al., Wüster et al., 1991; et al., 1992; Rosen et al., Beshyah et al., mass, assessed using several is greater than in body controls et al., 1992; et al., 1992; Rosen et al., Beshyah et al., et al., 1995). The increase in mass is observed in hypopituitary patients compared with controls et al., 1996). In increased has been demonstrated in hypopituitary adults by such of the ratio in several studies et al., 1992; et al., 1992; Cuneo et al., 1993; Rosen et al., Beshyah et al., et al., 1995). Increased and has also been confirmed using Beshyah et al., et al., 1995) and et al., 1995). hypopituitarism in man is a of increased 1991; 1992). The studies of et al. from our own demonstrated a more plasma and a in from following in GH deficient et al., and prolonged in and adults (Merimee et al., et al., GH the plasma et al., hypopituitarism is more associated with was similar in frequency in hypopituitary patients and the population in two series et al., Rosen et al., tolerance was found in of adults previously for in one of the series et al., Using the for the tolerance test observed a higher frequency of and impaired tolerance in our hypopituitary patients than in matched controls, particularly in women and in subjects (Beshyah et al., plasma concentrations were in hypopituitary adults in with in some studies et al., 1993; et al., 1995) but not significantly from and body mass controls in others (Beshyah et al., et al., 1994; et al., 1994). plasma concentrations were lower in our patients < than in the controls (Beshyah et al., plasma correlated significantly with and et al., 1994; et al., with body mass et al., 1993; and with body mass et al., 1995). the patients had higher and levels than matched controls (Beshyah et al., Using the Assessment was reduced to of normal and function was to et al., 1995). Using for and have demonstrated increased levels of circulating in hypopituitary patients with impaired tolerance (Beshyah et al., In to the of et al. in et al. (1995) using the tolerance test reported a significantly higher plasma in hypopituitary patients either or both and GH than in matched controls. The index was also significantly higher in the of and be to to the increased vascular In there may be an increase in serum and particularly in the of There is an increase in serum concentrations and normal or decreased et al., 1979; et al., 1979; et al., et al., et al., routine replacement abnormalities with pituitary disease on routine replacement have an serum (Merimee et al., and/or et al., 1979). In adult patients, Wüster et al. (1990) reported that had abnormal lipids and 42% and this was significantly higher than the population in the Rosen et al. showed that plasma levels were higher < 0.0001) than in the Swedish population but the in plasma were not statistically A higher prevalence of hypopituitary patients than in the population has been demonstrated in two et al., 1993; 1994). of hypopituitary patients and normal controls matched for and body mass including measurements of plasma and from 3 series are in levels were mainly due to increased levels of plasma and low may also concentrations of and higher were observed. plasma a vascular in one series the abnormalities were more in women (Beshyah et al., abnormalities were also in another series men only et al., 1994). The into the period and again the abnormalities were more in women than in men et al., 1997). low and concentrations in hypopituitary adults and matched controls are from 3 Cuneo et al., 1993; et al., and Beshyah et al., A to has been in clinical groups to be at risk of vascular including patients with and In hypopituitary patients the data are et al. (1990) showed normal levels of plasma and in mainly of childhood time and were normal in hypopituitary adults compared with matched controls (Beshyah et al., However, et al. (1994) 20 hypopituitary adults and 20 matched controls demonstrated higher plasma ± vs ± P < 0.0001) and ± vs ± P = in the patients, of a It is of to that acromegaly is associated with plasma levels are with et al., 1997). The results from different groups may in the patients studied although this is not from the A of is an important vascular risk In our studies there was no difference between patients and controls (Markussis et al., 1992). Hypopituitary patients either to the controls (Markussis et al., or than the population et al., is limited in adults with hypopituitarism and mass and are of of and low may although there are no data to The consequences for cardiovascular morbidity and mortality of GH therapy are at However, may be possible to from the and term effects on the cardiovascular and in cardiovascular risk factors. The cardiovascular effects of GH therapy in adults are in studies have demonstrated no or a effect on heart rate and is a of GH effect has been in animals, man and GH deficient of the in GH deficient et al., and in normal adults & 1990). However, in patients who had and there is evidence for both GH and IGF-I on the effect of GH on of has been to to et al., 1998). in all but one in an increase in left ventricular diastolic and and in left ventricular mass, in some but not in all studies. diastolic function was assessed in five studies and in at least one of diastolic function was demonstrated in two of The consequences of more prolonged are at are the effects of GH on the In GHD there was no significant in the body in all but one of the published studies et al., 1995). A significant increase in fat-free mass and a reduction in body has been demonstrated using several et al., et al., et al., 1992; Bengtsson et al., 1993; Rosen et al., 1994; Beshyah et al., et al., 1995; et al., Cuneo et al., 1998). body decreased significantly in some et al., Bengtsson et al., 1993; Beshyah et al., et al., 1995; Cuneo et al., 1998) but not in all studies et al., 1992; et al., 1996). Using Bengtsson et al. demonstrated that GH in a greater of than or body mass by (Beshyah et al., et al., 1995; et al., 1996; Cuneo et al., 1998) and from et al., 1995) decreased significantly during GH A similar effect of GH was demonstrated using in who are not et al., 1997). changes in body particularly the decrease in with GH are to be in of cardiovascular The effect of GH on is an The metabolic effects of GH replacement therapy in several are in GH for a significant but increase in in some but not other studies (Table There was a increase in et al., et al., 1994; Beshyah et al., significant was observed in et al., was there in et al., et al., or serum (Beshyah et al., during GH for a but in plasma and levels in response to in one study (Beshyah et al., but had no effect in another et al., 1992). from different decreased during GH replacement in some studies using et al., 1995) or tolerance test analysed by the of et al., but to a period of in other studies the et al., 1993) or the test et al., 1997). A significant reduction in serum and was reported GH in most but not all the studies (Table was observed in serum increased in some of the studies but this has not been However, an vascular risk increased on GH therapy in some of the GHD adults (Table et al., 1996). GH administration to normal subjects is also associated with an increase in plasma et al., is and with serum GH levels in acromegaly et al., that GH is an important of of GH does not to be by administration of IGF-I decreases concentrations et al., 1995). here are limited and are not et al. (1990) reported changes in factors of GH replacement = There were significant in plasma and and in and showed some to All in the range and the of the was by the results from another of 20 subjects a decrease in plasma = an increase in = and a significant increase in et al., 1994). In a smaller study of 8 patients for 8 no significant changes were observed in time et al., More the effects of term of GH therapy on and were studied in patients with adult-onset GHD et al., 1996). concentrations to decrease = the factors and were and demonstrated a decrease of GH The decreased < was and levels decreased during the GH period < The decrease was more in patients with levels of the different decreased < was during two of GH increased vascular mortality has been to GH deficiency & Wüster et al., there does not to be a for all the abnormalities the clinical of hypopituitarism in adults. of other pituitary may to the increased vascular replacement with or could changes and present replacement are not particularly to the abnormalities of and GH deficiency may be the underlying the abnormal body composition and and may also to the and be to effects of GH deficiency. However, GH deficiency may in to these abnormalities the abnormal body although some abnormalities such with GH replacement in body long-term GH the cardiovascular risk the between its different The effect of GH may be by the effect for of the increase in and decrease in and available data is not of the effects on tolerance and to in some of the et al., Bengtsson et al., 1993; et al., 1997) but not in others et al., 1994; Beshyah et al., et al., 1995). excess is associated with increased and body It is also associated with and et al., et al., concentrations are et al., also causes from vascular disease et al., in hypopituitary is therefore a potential to vascular replacement is most It has the of and most patients the normal in peak between and et al., et al., 1978; et al., with the are low or et al., 1996). patients are for to time but the with low this some a such for does not of the of and the in the of There is no of to the of therapy & & 1992). In some the is by clinical others and in other are is most data during the and replacement at in both adults and & et al., 1997). and levels during were higher of than levels on when replacement therapy was the test et al., 1995). The of this for long-term vascular risk is However, has been that with may be an important to the observed excess cardiovascular disease 1994; 1997). GH replacement has effects on et al., 1994; et al., 1996; et al., et al., 1998). The effects were only in patients replacement GH reduced serum serum decreased the of the of The ratio of of to those of was reduced in one study et al., but not in other two et al., et al., 1998). However, the ratio of to that of increased in one of the studies et al., 1998). It was that of may be reduced by and that GH the and The clinical significance of these are not However, of in the of patients replacement therapy may to the effect of GH et al., 1998). The of these for the GH deficient are deficiency the of atheroma in the in man causes an in serum and is therefore associated with excess vascular risk and with risk factors. a normal circulating hormone with an the of pituitary causes and an increased cardiovascular risk et al., The with replacement at least in some studies. In patients with the is not a to the of hormone replacement. of serum hormone concentrations has therefore been all the of hormone normal a It is therefore to be that patients are though circulating hormone levels may be in the normal may also to vascular risk and risk factors. have important and on the cardiovascular and the cardiovascular risk factors et al., 1995). hormone replacement is mostly and on and therefore or replacement is A significant to of hypopituitary women in series were not replacement The mortality in hypopituitary patients were higher in women than in men in all the published studies vs & vs (Bates et al., & 2.39 vs (Bulow et al. hypopituitary women had more abnormalities of the carotid wall structure and function (Markussis et al., 1997) and body and metabolic abnormalities such and hyperlipidaemia (Beshyah et al., 1994; Excess body similar to that to GH deficiency, has been demonstrated in women at the et al., and has been to be by replacement therapy et al., The in vascular risk a possible from hormone replacement in these groups of patients In metabolic of other than are not although these are in other In a study of 8 subjects with of levels ± there was higher plasma levels and smaller of both during tolerance compared with matched controls for and body et al., 1995). In another study et al., body and were found to be associated with pituitary levels & ± in patients with = and = respectively). The was of mass effect and other pituitary in of the patients of of and of plasma of may be by patients with pituitary disease although levels are The significance of in adults and its possible role in increased vascular risk