| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Service de Pharmacologie Clinique (P.M., A.-I.T., I.M.-M.), Unité de Recherche Clinique (P.M.), Assistance Publique-Hôpitaux de Paris, Hôpital Henri Mondor, and University of Paris 12 (P.M.), F-94010 Créteil, France; Institut National de la Santé et de la Recherche Médicale U421 (P.M.), F-94004 Créteil, France; Department of Internal Medicine (A.G.), University of Brescia, 25125 Brescia, Italy; and Service d’Endocrinologie et des Maladies de la Reproduction (P.C.), Assistance Publique-Hôpitaux de Paris, Hôpital de Bicêtre, University of Paris Sud 11 (P.C.), and Institut National de la Santé et de la Recherche Médicale U693 (P.C.), F-94276 Le Kremlin-Bicêtre, France
Address all correspondence and requests for reprints to: Professor P. Chanson, M.D., Service d’Endocrinologie et des Maladies de la Reproduction, Hôpital de Bicêtre, 78 Rue du Général Leclerc, Le Kremlin-Bicêtre, F-94275, France. E-mail: philippe.chanson{at}bct.ap-hop-paris.fr.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Objective: The objective of the study was to conduct a metaanalysis aimed at obtaining a more accurate picture of the effect of somatostatin analogs on the heart in patients with acromegaly.
Design: We systematically reviewed all studies of somatostatin analogs in acromegaly. Eighteen studies were identified in three databases. We conducted a combined analysis of the effects of somatostatin analogs by using the overall effect size to evaluate significance and by computing the weighted mean differences with and without treatment to assess the effect size.
Results: Somatostatin analog treatment was associated with significant reductions in the heart rate [–5.8 (2.1) beats/min], the left ventricular mass index [–22.3 (6.7) g/m2], interventricular septum thickness [–0.3 (0.2) mm], left ventricular posterior wall thickness [–0.8 (0.4) mm], and the ratio of the E-wave and A-wave peak velocities of the mitral flow profile [0.2 (0.1)]. It was also associated with improved exercise tolerance [1.6 (0.4) min]. Trends toward beneficial effects were noted for the left ventricular end-diastolic dimension [–1.5 (2.2) mm] and the left ventricular ejection fraction [3.3% (1.7%)]. Overall effect sizes were not significant for blood pressure, left ventricular end-systolic dimension, or fractional shortening. Bigger improvements were observed in studies with larger falls in IGF-I and/or GH levels and studies of younger patients.
Conclusion: This metaanalysis confirms that somatostatin analog therapy aimed at achieving stringent control of serum GH/IGF-I concentrations in patients with acromegaly is associated with significant positive effects on morphological and functional hemodynamic parameters.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Previous studies addressing the reversibility of cardiomyopathy in treated acromegalic patients gave variable results (reviewed in Ref. 16).
The advent of somatostatin analogs, which are capable of rapidly normalizing GH/IGF-I levels in a majority of patients, has led to a reappraisal of this issue. The effects of somatostatin analogs have been evaluated in terms of not only GH/IGF-I suppression but also their impact on morbidity and particularly cardiomyopathy. Improvements in cardiac parameters obtained with somatostatin analogs (5) have been reported, probably owing to better GH/IGF-I control. However, most studies involving somatostatin analogs involved small numbers of patients, raising the possibility that nonsignificant results were related to inadequate statistical power. We therefore conducted a metaanalysis to obtain a more accurate picture of the impact of somatostatin analogs on the heart in patients with acromegaly.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We searched the medical literature for reports on the cardiac effects of somatostatin analogs in acromegaly. We searched three electronic databases, Medline (Ovid), Biosis, and Experta Medica (EMBASE), from their year of inception to June 2006. We used a combination of the following key words: lanreotide, octreotide, somatostatin analog, ventricular, heart, and acromegaly. The search strategy was not limited by study design or language. Abstracts from major cardiology meetings were examined in an attempt to identify unpublished trials and thereby reduce the risk of a publication bias.
Inclusion criteria
We first included all studies of somatostatin analogs in acromegaly. The selected publications had to report at least one of the following outcome measures: heart rate (HR), interventricular septum (IVS), left ventricular (LV) posterior wall (LVPW), left ventricular end-diastolic dimension (LVEDD), LV end-systolic diameter (LVESD), LV mass (LVM), LV mass index (per square meter of body surface area) (LVMi), left ventricular ejection fraction (EF), ratio of early to late mitral diastolic flow (E/A), and exercise duration.
Data extraction
Data were extracted from published reports onto standardized forms by two metaanalysts (A.-I.T. and P.M.). Discrepancies were resolved by discussion among the authors of the present paper. The following data were extracted: mean age, sex, number of patients included, duration of acromegaly, conventional heart failure therapy, somatostatin analog name, target dose, treatment duration, GH concentration at baseline, IGF-I decline during treatment, study quality (design, blinding, statistical methods), losses to follow-up for each outcome measure, and baseline and follow-up values in comparator groups (means and SD or SEM).
Statistical methods
For primary analyses of continuous outcome measures, we calculated standardized effect sizes for each trial and global effect sizes for each outcome. In the study by Lombardi et al. (17), the values had to be estimated from the figures. The effect sizes were calculated by using the mean difference between baseline and follow-up divided by the variance at baseline. A positive effect size is an increase during somatostatin analog treatment, and a negative effect size is a decrease. To calculate the overall effect size, each study was weighted by the reciprocal of its variance. When the variance of changes was not stated, it was calculated from the confidence intervals (variances, SEM) (18). We report the overall effect sizes with their 95% confidence intervals.
We explored heterogeneity across studies with the Q test. When the effect size was significant in a fixed model but the Q test was significant, the analyses were repeated with a random-effects model to confirm the result. In these cases, we report the effect sizes according to the random-effects model.
Funnel plots were drawn and their asymmetry was measured to assess the possible influence of publication and location biases (19). The intercept of linear regression, in which the effect size divided by the SE is regressed against the reciprocal of the SE, provides a measure of asymmetry. To assess sensitivity, when the effect size was dependent on one or two trials (e.g. a large trial), these trials were dropped from the analysis.
To quantify the effect size, we calculated the weighted mean (and SD) of the change between the periods, for each outcome measure, when the data were available.
The effects of the baseline GH concentration, the IGF-I decline, disease and treatment durations, patient age, and the degree of LV dysfunction on the overall estimates were assessed by stratification or metaregression. Weighted least-squares regression was used for metaregression, with individual study effects weighted by the reciprocal of the estimated variance. The ß coefficient and its significance along with the adjusted R2 value are reported to show the overall variability explained by the model.
Analyses were conducted with the SPSS 13.0 (SPSS Inc., Chicago, IL) package for Windows.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). Octreotide sc, octreotide long-acting release (LAR) and lanreotide prolonged release (PR) were used in 10, five, and three studies, respectively. Mean (SD) GH levels (n = 16 studies) and IGF-I levels (n = 15) were, respectively, 35 (20) and 723 (196) µg/liter at baseline and 7.6 (8.2) and 406 (148) µg/liter at follow-up. The studies were generally of good quality, as shown by the small number of losses to follow-up [subjects were lost to follow-up in only one study (23)] and by the statistical methods used. None of the studies were randomized controlled trials, however.
|
). Funnel plots (not shown) and linear regression did not suggest a selection bias, except for LVMi (ß = 0.03; P = 0.002).
|
) with respect to HR [–0.47 (–0.74; –0.19)], LVM [–1.14 (–2.06; –0.21)], LVMi [–0.78 (–0.97; –0.58)], IVS [–0.67 (–0.95; –0.38)], LVPW [–0.50 (–0.87; –0.14)], E/A [0.43 (0.16; 0.70)], and exercise duration [1.06 (0.43; 1.70)]. A trend was noted for LVEDD [–0.20 (–0.43; 0.03)] and EF [0.34 (–0.08; 0.76)], which were measured in eight and 14 studies, respectively. Overall effect sizes were not significant for BP, LVESD, or FS (Table 2). Because a selection bias was suspected for LVMi, effect sizes were recalculated after excluding the study responsible for asymmetry (32). The global effect sizes remained significant [–0.85 (–1.06; –0.64)].
The weighted mean change (and SD) before and after treatment with somatostatin analogs was determined for each outcome measure to quantify the size of the effect (Table 2).
When the analysis was confined to studies of octreotide sc, the significant effect sizes were confirmed for HR [–0.43 (–0.71; –0.15)], LVMi [–0.58 (–0.94; –0.21)], IVS [–0.72 (–1.03; –0.41)], E/A [0.52 (0.00; 1.04)], EF [0.24 (0.01; 0.48)], and exercise duration [0.57 (0.15; 0.98)]. With octreotide LAR, HR [–0.73 (–1.33; –0.13)], LVMi [–1.04 (–1.35; –0.73)], EF [0.36 (0.05; 0.68)], and exercise duration [1.43 (0.52; 2.33)] were significant. Similar results were obtained with lanreotide PR for LVMi [–0.88 (–1.33; –0.43)], IVS [–0.44 (–0.76; –0.13)], LVPW [–5.7 (–7.46; –3.98)], E/A [0.40 (0.09; 0.71)], and exercise duration [2.24 (1.63; 2.84)].
In the subgroup analysis based on median age, significant effect sizes were observed for PW and LVM among patients under 45 yr. Global effect sizes showed significantly bigger improvements in IVS [–0.94 (–1.83; –0.05)] and exercise duration [1.53 (1.06; 1.99)] in studies with younger vs. older patients [respectively, –0.54 (–0.87; –0.02) and 0.47 (0.12; 0.81)]. Although these results suggest an age/effect relationship, no firm conclusion could be drawn because of the small number of studies.
When enough studies were available, we conducted subgroup analyses based on the median decline in IGF-I (50%) during treatment. Bigger improvements were observed in studies with larger declines in IGF-I levels (>50%, as compared with <50%) with respect to IVS [–1.09 (–1.51; –0.68) vs. –0.41 (–0.70; –0.13)], LVPW [–0.91 (–1.49; –0.34) vs. –0.15 (–0.86; 0.57)], and LVM [–2.00 (–3.41; –0.60) vs. –0.41 (–1.37; 0.55)]. Subgroup analyses based on the median decline in GH levels confirmed these results, with bigger improvements in studies with larger declines in GH (>50%, compared with <50%) with respect to exercise duration [1.48 (0.40; 2.57) vs. 0.51 (0.07; 0.94)] and LVPW [–0.84 (–1.22; –0.46) vs. 0.10 (–0.32; 0.52)].
Bigger improvements were observed in studies with low baseline LVM values (less than median, compared with greater than median) with respect to LVMi [–1.04 (–1.33; –0.74) vs. –0.58 (–0.83; –0.33)] and exercise duration [1.67 (1.03; 2.30) vs. 0.33 (–0.06; 0.72)]. No significant difference was observed according to BP or gender. There were too few studies for secondary analyses based on the duration of the disease or treatment.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
As in all systematic overviews, nonpublication is potentially a major source of bias because trials with positive and significant results are more likely to be published than those with neutral or negative results. Nevertheless, because the significant parameters were rarely identical in the selected trials, such a bias seems unlikely. The main limitation of this metaanalysis was the lack of randomized control trials.
One of the most distinctive features of acromegalic heart involvement is LV hypertrophy (4, 10, 34). Our metaanalysis shows a significant positive effect of somatostatin analogs on LVM. This is likely due to the concomitant decrease in LVPW, confirming the results of most published trials (17, 22, 29, 30), and to the decrease in IVS (17, 21, 22, 23, 27, 29, 30), rather than to the fall in LVEDD found in one study (29). Cardiac hypertrophy in acromegaly is at least partly independent of arterial hypertension and is linked to the hyperactivity of the GH/IGF-I axis. Acromegaly is associated with hypertension in nearly 50% of patients (35), and potential pathogenetic mechanisms include sodium-fluid retention (36), peripheral vasomotor dysfunction (37, 38), and endothelial disturbances (10, 39, 40). Our metaanalysis showed no clear reduction in BP during somatostatin analog therapy. Although one study showed an effect of somatostatin analogs on hypertension (5), our analysis supports a pressure-independent effect of somatostatin analogs on heart mass. The mechanism of action probably involves suppression of GH/IGF-I hypersecretion, even if a direct positive effect of these drugs on cardiomyocytes cannot be ruled out (41).
This reduction in ventricular hypertrophy could have both prognostic significance and functional benefits. Indeed, ventricular hypertrophy has an independent negative prognostic role in essential hypertension (42, 43). In addition, the reduction in ventricular hypertrophy might, in theory, improve diastolic dysfunction. Diastolic function, assessed in terms of the ratio of E-wave to A-wave peak velocities of the mitral flow profile, improved in only one published study (32), whereas a trend toward an improvement was observed in another study (21). Our metaanalysis nonetheless confirms these results. The somatostatin analog-induced reduction in the heart rate demonstrated in our metaanalysis may also contribute to this improvement. It is likely to occur via an effect on autonomic tone, which may superimpose on the sympathovagal imbalance due to vagal hypertone found in untreated patients with acromegaly (44).
Only two studies showed a significant effect on the ejection fraction, which reflects systolic function (25, 28). The ability of somatostatin analogs to increase contractility is controversial. One study showed an increase in fractional shortening (25), whereas others showed no change (20, 22, 23, 27). Our metaanalysis showed no positive effect on fractional shortening. It must be kept in mind, however, that improved LV FS reflects an increase in contractility only if loading characteristics remain unchanged. Most studies showed that postloading conditions (LVESD) were not improved by somatostatin analogs (20, 22, 24, 25, 27, 29). Nevertheless, a trend toward an increase in the ejection fraction was seen in our metaanalysis. We also confirm the beneficial effect on exercise duration observed in five studies (24, 26, 31, 33). Overall, this seems to suggest that somatostatin analogs may lead to an overall improvement in cardiac performance in acromegaly.
It has been claimed that stringent control of GH/IGF-I concentrations, as defined by consensus criteria (14, 15), bolsters the improvement in cardiac parameters. We were unable to compare controlled and uncontrolled patients in our metaanalysis owing to a lack of individual data in most studies. However, we confirm that IVS, LVPW, and LVM tended to improve more in studies with larger decreases in serum IGF-I. This was also true for exercise duration and LVPW in studies with larger decreases in GH.
Most parameters differed widely among the studies, possibly owing to differences in patient populations. Nevertheless, our analysis supports differences in the treatment effects according to age and the degree of hypertrophy (33). Larger trials or metaanalysis of individual data is needed to explore these sources of variability and their interactions.
In conclusion, this metaanalysis confirms that somatostatin analog treatment targeting stringent control of GH/IGF-I concentrations has a significant positive effect on LVM, LVPW thickness, LVEDD, E/A, and EF (as assessed by echocardiography), and on exercise duration, in patients with acromegaly.
|
Footnotes |
|---|
First Published Online February 20, 2007
Abbreviations: BP, Blood pressure; E/A, ratio of early to late mitral diastolic flow; EF, ejection fraction; FS, fractional shortening; HR, heart rate; IVS, interventricular septum; LAR, long-acting release; LV, left ventricular; LVEDD, LV end-diastolic dimension; LVESD, LV end-systolic diameter; LVM, LV mass; LVMi, LV mass index (per square meter of body surface area); LVPW, LV posterior wall; PR, prolonged release.
Received November 20, 2006.
Accepted February 13, 2007.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  Y. Hua, H. Ma, W. K. Samson, and J. Ren Neuronostatin inhibits cardiac contractile function via a protein kinase A- and JNK-dependent mechanism in murine hearts Am J Physiol Regulatory Integrative Comp Physiol, September 1, 2009; 297(3): R682 - R689. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Giustina, S. Bonadonna, G. Bugari, A. Colao, R. Cozzi, S. Cannavo, L. de Marinis, E. degli Uberti, F. Bogazzi, G. Mazziotti, et al. High-dose intramuscular octreotide in patients with acromegaly inadequately controlled on conventional somatostatin analogue therapy: a randomised controlled trial Eur. J. Endocrinol., August 1, 2009; 161(2): 331 - 338. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Mazziotti, I. Floriani, S. Bonadonna, V. Torri, P. Chanson, and A. Giustina Effects of Somatostatin Analogs on Glucose Homeostasis: A Metaanalysis of Acromegaly Studies J. Clin. Endocrinol. Metab., May 1, 2009; 94(5): 1500 - 1508. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Melmed, A. Colao, A. Barkan, M. Molitch, A. B. Grossman, D. Kleinberg, D. Clemmons, P. Chanson, E. Laws, J. Schlechte, et al. Guidelines for Acromegaly Management: An Update J. Clin. Endocrinol. Metab., May 1, 2009; 94(5): 1509 - 1517. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. A van der Klaauw, J. J Bax, G. B Bleeker, E. R Holman, V Delgado, J. W A Smit, J. A Romijn, and A. M Pereira Cardiac manifestations of GH deficiency after treatment for acromegaly: a comparison to patients with biochemical remission and controls Eur. J. Endocrinol., December 1, 2008; 159(6): 705 - 712. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Mazziotti, A. Bianchi, S. Bonadonna, V. Cimino, I. Patelli, A. Fusco, A. Pontecorvi, L. De Marinis, and A. Giustina Prevalence of Vertebral Fractures in Men with Acromegaly J. Clin. Endocrinol. Metab., December 1, 2008; 93(12): 4649 - 4655. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Giustina, G. Mazziotti, and E. Canalis Growth Hormone, Insulin-Like Growth Factors, and the Skeleton Endocr. Rev., August 1, 2008; 29(5): 535 - 559. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
