Diabetes in acromegaly, prevalence, risk factors and evolution; data from the French acromegaly

Transcription

1 Page 1 of 27 Accepted Preprint first posted on 4 April 2011 as Manuscript EJE Title: Diabetes in acromegaly, prevalence, risk factors and evolution; data from the French acromegaly registry. Authors: Sandrine Fieffe (1), Isabelle Morange (2), Patrick Petrossians (3), Philippe Chanson (4), Vincent Rohmer (5), Christine Cortet (6), Françoise Borson-Chazot (7) Thierry Brue (2), Brigitte Delemer (1) and the French Acromegaly Registry. 1. Service d Endocrinologie, CHU de Reims, 51092, Reims France. 2. Service d Endocrinologie, Diabète, et Maladies Métaboliques, et Centre de Reference des Maladies Rares d Origine Hypophysaires, Hôpital de la Timone, F Marseille, France 3. Service d Endocrinologie, CHU de Liège, Domaine Universitaire du Sart Tilman, 4020 Liège, Belgium 4. Service d Endocrinologie, CHU de Bicêtre, Le Kremlin-Bicêtre, France. 5. Service d Endocrinologie, CHU d Angers, 49033,Angers, France. 6. Service d Endocrinologie, Clinique Marc Linquette,59037, Lille, France. 7. Service d Endocrinologie, Hôpital de l Antiquaille, 69005, Lyon France. Corresponding Author: Professeur Brigitte Delemer, Endocrinologie,Hôpital Robert Debre, rue du General Koenig, 51090, Reims,France Running title: Diabetes in acromegaly. Copyright 2011 European Society of Endocrinology.

2 Page 2 of 27 Abstract: Objectives: The French Acromegaly Registry records acromegalic patients data since 1992, in French, one Belgian (Liège) and one Swiss (Lausanne) centers. We studied the prevalence of diabetes in this population looking for risk factors. Patients from one of the centers (Reims) were then analyzed more thoroughly. Methods: This study has been conducted on all the patients recorded from 1999 until 2004 (519 patients). Cohorts evolution was reassessed in Of the different variables recorded in the registry, age, sex, BMI, duration of acromegaly, GH level, IGF-1 and prolactin levels, pituitary tumor size, hormonal deficiencies, presence, duration and treatment of diabetes, hypertension and rheumatological disease were analyzed. Results: Prevalence of diabetes in the registry was 22.3%. Diabetic patients were older and had a higher BMI. Compared with the data of the french social security, acromegalic patients showed a more precocious apparition of diabetes and prevalence was higher in each age group. Compared to non-diabetic acromegalic subjects, diabetic patients had a more prolonged evolution of acromegaly before diagnosis. Levels of GH and IGF-1 were not significantly different between the 2 groups. Only hypertension was significantly more frequent in diabetic patients. Conclusions: In our population, the prevalence of diabetes was estimated to be 22.3%. Growth hormone and IGF- 1 levels did not appear as predictive factors for the presence of diabetes. In contrast, age, BMI and hypertension were significant risk factors as in the general population of type 2 diabetics.

3 Page 3 of 27 1 Introduction Acromegaly is a disease characterized by growth hormone (GH) and insulin-like growth factor 1(IGF-1) hypersecretion, due, in most cases, to a pituitary somatotropic adenoma [1]. The diagnosis of acromegaly is usually delayed for years, exposing patients to slowly evolving chronic complications [2]. This disease leads to increased morbidity and mortality linked in the first place to cardiovascular disease [3, 4, 5]. Excess of GH and IGF-1 interacts with metabolic regulation and indeed, GH hypersecretion is associated with hepatic and peripheral insulin resistance [6] this, and also other mechanisms lead to the development of diabetes mellitus. Hyperinsulinism, insulin resistance and diabetes are well recognized cardiovascular risk factors in general population and may participate as well to the increased cardiac morbidity and mortality of acromegalic patients [7, 8, 4, 5]. Incidence of diabetes is diversely evaluated in clinical series of acromegalic patients because of insufficient number of patients in the different cohorts [2]. Since several years, studies of rare diseases have been facilitated by construction of large registries in different countries [9]. In this paper we report data from the French Acromegaly Registry, which collects clinical data from acromegalic patients diagnosed since 1993 in participating french-speaking centers. The aim of our study was to precisely establish the prevalence of diabetes in this cohort and to search favoring factors for the presence of diabetes in such patients. Furthermore, we studied long term evolution of

4 Page 4 of 27 diabetes after initiating the treatment of acromegaly. As the collection of the data in the registry was not exhaustive, we analyzed with more depth the data of a regional study led in one of the participating centers (Reims) in which all the consecutive patients diagnosed since 1993 have been included. 2 Patients and methods 2.1 Patients The French Registry of Acromegaly has been initiated in 1999 by the French Society of Endocrinology. Data from patients diagnosed with acromegaly since 1993 in participating centers were recorded in a central database. Written informed consent was obtained from all subjects. Patient data were grouped in different visits. Initial data were included in a so-called historical visit, data at time of inclusion in an inclusion visit, and annual follow-up data were recorded in follow-up visits. This study has been conducted on all the data recorded from January 1999 until July 31st 2004 and the evolution of the cohort has been analyzed in November Methods Diagnosis of acromegaly was based on Cortina criteria [10]:either nadir of GH > 1 ng/ml on oral glucose tolerance test (OGTT) or, especially for diabetic patients, mean GH > 2.5 ng/ml and IGF-1 levels superior to upper limit of normal for age and sex. Diabetes mellitus was diagnosed using WHO s 1998 criteria [11]. Patients of 15 years old or less were excluded from the study as evaluation of Body Mass Index (BMI) and Blood Pressure (BP) are different in childhood. For each patient, we analyzed the following data recorded by the local investigator at the diagnosis of acromegaly : age and sex, BMI, estimated duration of acromegaly, mean plasmatic GH level, IGF-1 level (absolute level and % of the upper limit of normal for age and sex in each center), prolactin levels, pituitary tumor size measured on MRI, presence of hormonal pituitary deficiencies. The following chronic complications were evaluated : presence, duration and treatment of diabetes,

5 Page 5 of 27 hypertension and rheumatological disease (arthralgia, carpal tunnel syndrome or manifestation of arthrosis). Treatment of diabetes have been graduated as followed : 0 for diet only, 1 for metformin, 2 for sulfonylureas, 3 other treatments and 4 insulin. Hormonal measurements were realized with commercial kit available in each center and the normal reference range at each time point was used. 2.3 Statistical analysis Quantitative variables cited in this article did not satisfy the normality test. These data were therefore expressed as mean and median value, first and third quartiles (Q1=percentile 25, Q3=percentile 75). For variables referred-in in the discussion, a density graph was plotted in order to allow a better approach of data distribution between the two groups. Comparisons in univariate statistics were made with the non-parametric Mann-Withney test. Count data were expressed as absolute value and percentage, comparison in univariate statistics were made with the chi-squared test. Risk factors were assed with multivariate regression logistic test with an entry threshold of <0.15 and an exit threshold of >0.05. Significance threshold was set at <0.05 for all analyses. The SAS software (Cary, USA) was used for the statistical analyses, graphics were designed with the R statistical software. 3 Results 3.1 Study population The study population is summarized in tables 1 and 2. The mean age of acromegalic patients in the registry was 46.1 yrs., with a predominance of female (male to female sex ratio of 0.8). Mean duration of acromegaly based on the assessment of endocrinologists at diagnosis was 7.5 years. MRI revealed a pituitary macroadenoma in 80% of patients and endocrinological assessment demonstrated an associated pituitary deficiencies in 26% of cases. Most frequently reported comorbidities were osteoarticular complications (58%) and hypertension (34%).

6 Page 6 of Prevalence of diabetes and characteristics of diabetic patients Prevalence of diabetes in the first 519 patients reported in the registry was 22.3%. When only patients treated by oral antidiabetics and/or insulin were taken into account (excluding patients under diet), the prevalence was 13%. Male to Female sex ratio in diabetic patients (0.8) was not significantly different compared to non diabetic patients (0.84, p=0.85). Age and BMI (fig. 1) were significantly different in diabetic patients who were older (p<0.001) and had a higher BMI (p=<0.001). However, there were precocious cases of diabetes with the earliest one occurring at 26 YO. In patients from the registry, there was a steady increase of percentage of diabetes with age. Compared with the data of the french social security (prevalence of treated diabetes [12]) or the INSTANT study (prevalence of type 2 diabetes in the french population [13]), acromegalic patients showed a more precocious apparition of diabetes and percentages were higher in each age group. Regarding acromegaly parameters, the only significant difference observed in diabetic patients vs non diabetic patients was a more prolonged evolution of acromegaly before diagnosis (p<0.001). Neither hormonal levels of GH (p = 0.18) nor IGF-1 (p = 0.4) were significantly different between the 2 groups. Diabetes was diagnosed at the same time as acromegaly in 46.8% of patients. For all other patients, the discovery of diabetes predated the diagnosis of acromegaly with a mean duration of 5.8 years and a median duration of 4 years. Although based on clinical estimation and patient anamnesis, diabetes was present before the presumed beginning of acromegaly in only 9.8% of patients. Diabetes was treated with only diet in 39.8% of patient, metformin was used alone in 13% of patients, sulfonylureas or sulfonylureas associated with metformin in 29.6% of patients and insulin in 17.6%. Hypertension was significantly more frequent in diabetic patients (61.5% vs 26.3%, p<0.001). This was not the case for rheumatological complications (p= 0.14) or pituitary deficiencies (p= 0.271).

7 Page 7 of Diabetes risk factors in acromegaly : We performed both univariate and multivariate analysis of recorded variables. Age at diagnosis, BMI, hypertension and duration of evolution of acromegaly were significant risk factor for diabetes in acromegalic patients upon univariate analysis. Age, BMI and HTA were independant risk factor upon multivariate analysis. Presence of hypertension increased the risk of diabetes by 2.5. For an acromegalic patient the risk to present diabetes was increased by 4.4% by year of age (Odd ratio, OR = 1.044) and by 12.9 % by additional kg/m2 (OR = 1.129) 3.4 Regional study in Reims In this center, all the consecutive acromegalic patients from 1993 until July 2004 were included. They were 74 patients among which 24 were diabetic (32%), significantly more than in the registry (p=0.04). But the patients were also older (mean age=50.2 yrs. in Reims vs 46.1 yrs. in the registry) and with a longer duration of acromegaly (mean duration=10.6 yrs. vs 7.5 yrs.). In this population the same risk factors for diabetes were found (age, BMI, hypertension, duration of acromegaly). Furthermore, we had the opportunity to study the presence of familial history of diabetes which is an additional risk factor for diabetes not available in the registry data. We found a significant increase of this parameter in diabetic acromegalic patients (p=0.05) 3.5 Follow-up data Follow-up of the 114 diabetic patients after treatment of acromegaly was studied in November Information was available in 108 cases after a mean evolution time of 5.40 yrs. (1-12 yrs.). Acromegaly was controlled in 67 patients (62%), 26 of whom were remaining under medical treatment. In these patients, diabetes had disappeared in 23 (30%) cases. When acromegaly was still active, diabetes was normalized in 9 out of 41 patients (22%). However, this trend did not appear to be statistically significant (p=0.25). Treatment score of patients with active acromegaly was higher, but not significant, when compared to patients who were cured or controlled (Mean:1.561 vs 1.299,

8 Page 8 of 27 p=0.12). In the cohort of still diabetic patients, treatment had been modified as follow : 29 patients were on diet only vs 24 at diagnosis of acromegaly, 17 patients on metformin vs 11, 18 patients on sulfonylureas vs 25 and 12 on insulin vs 16 patients. Insulin treatment has been stopped in 8 patients over 16 but had been initiated in 4 other patients. We looked for predictive factors of remission of diabetes. We evaluated age, GH and IGF-1 levels at diagnosis, remission of acromegaly after treatment, duration of diabetes, treatment of diabetes. We found that remission of diabetes seemed likely if patients were younger (p=0.075) with a more recent diabetes (p=0.087) treated by diet only (p=0.013), although for the first two variables, p-values were higher than the 0.05 significance value. However, there were some exceptions. In patients whose diabetes was cured we found 3 patients with a diabetes of more than 10 years duration, 3 patients who were treated by insulin and 8 patients whose acromegaly was not controlled but with a dramatic improvement of GH. We also compared the evolution of diabetes in GH and IGF-1 «dissociated» patients (patients who normalized GH but not IGF-1 vs patients with normalized IGF-1 but not GH) without finding significant results (respectively p=0.31 and p=0.36). During this follow-up period, de novo diabetes was diagnosed in only one patient. 4 Discussion This work is based on the French Registry of Acromegaly that collects information on acromegalic patients diagnosed since 1993 in French, Belgian and Swiss tertiary centers to improve knowledge of this rare disease and of its outcome in a real life setting. We collected data about diagnosis, associated morbidities and their evolution, treatments and their results. In this publication, we focus our interest on diabetes which is known to be a frequent complication of acromegaly [2]. Glucose metabolism was evaluated in all of Registry s patients as OGTT was considered mandatory for inclusion and our cohort of 519 patients seemed sufficient to evaluate such a frequent complication. At present, this study constitutes the most comprehensive series of cases recorded for evaluation of diabetes in de novo acromegalic patients although it is not an exhaustive collection of cases during

9 Page 9 of 27 the study period ( ) as it represents only a fraction of what we could have expected according to the recently reevaluated prevalence of acromegaly by Daly et al. [14]. The Belgian study describes a prevalence of 120 cases of acromegaly per million of inhabitants, which, reported to the French population, would predict a population of more or less 6000 acromegalic patients. We controlled our work by a single center study in which all the consecutive cases of acromegaly recorded between 1993 and 2004 have been included. Our local results agreed with the results of the whole cohort. Characteristics of acromegalic patients followed in the french registry are similar to what is reported in other series or registries considering mean age (46.1 yrs.) as well as discrete female predominance (M/F sex ratio 0.8) or mean duration of active acromegaly before diagnosis (7.5 years) [15, 16, 17, 18]. Complications are frequent and are the consequences of a delayed diagnosis : 34% of our patients presented with arterial hypertension and 58% with rheumatological diseases. Tumors were macroadenomas in more than 80% of patients and pituitary deficiencies were present in 26% of the cases. The prevalence of diabetes mellitus among patients with acromegaly in the french registry is 22.3%. Results from other series reporting more than 100 patients are shown in table 3. Our work gives a higher prevalence of diabetes compared to the older studies may be because of the modification of the criteria for the diagnosis of diabetes since The Spanish [19] and the Belgian registries [20] found a still higher incidence of respectively 37.6 and 25.3%. However, these two series are very different from ours as their objective was to study all the known acromegalic patient followed in a defined territory. Some of the cases had been diagnosed a number of years ago. Furthermore, this complication was not evaluated specifically at the diagnosis and could concern older patients with a longer evolution time. To determine the impact of acromegaly in the prevalence of diabetes we compared our results to the data obtained in 2005 in the general population by Social Security in France (treated diabetes) [12] and the results of the INSTANT study (prevalence of diagnosed diabetes in a representative sample, either treated or not) [13]. Global prevalence of diabetic patients treated by oral antidiabetic drugs or insulin was 3.6% in Social Security data, 4.5% in the INSTANT study and 13% in our cohort. When results

10 Page 10 of 27 were stratified by age and gender, acromegaly increased the incidence of diabetes particularly in younger people (30 to 50 Y.O.) and in women. Diabetes was diagnosed in half of the cases at the same time as acromegaly. In only 9.6% of the cases diabetes preceded acromegaly and only one patient developed diabetes during follow-up. Furthermore, after treatment of acromegaly, diabetes disappeared in 30% of the cases and treatment could be reduced in a number of the other cases. The likelihood to control diabetes increases if acromegaly is also controlled. In another study evaluating comorbidities in strictly controlled acromegalic patients, prevalence of diabetes was not different (11%) from a large control group (8.5%). All these results confirm the large implication of GH excess for the presence of diabetes. However, we could not identify GH nor IGF-1 levels as significant risks factors for diabetes. Variability in different commercially available kits for measuring GH and IGF-1 may have blurred the role of hormone concentrations in the presence of diabetes. Another way to approach the impact of GH excess is to estimate the duration of acromegaly although it is a clinical parameter which is very difficult to assess. We found that diabetic patients had a significantly longer evolution of acromegaly but disease duration did not appear as an independent variable on multivariate analysis. Although not directly related to acromegaly, we were able to demonstrate 3 independent parameters: age, BMI, and hypertension. In our local study, familial history of diabetes appeared as a possible additional factor. These parameters have nothing to do with acromegaly per se but are rather related to what is known about diabetes risk factors in the general population. Few other authors have addressed this question. Wass et al analyzed the data of 69 acromegalic patients and did not find any parameter in relation with occurrence of diabetes [21]. Nabarro et al reported the results of a large cohort of 208 patients among which 48 had diabetes [17]. They found a significant correlation between age, duration of evolution of acromegaly and GH level with the presence of diabetes but only by univariate analysis because of the insufficient size of the cohort. Finally, Biering et al identified age but not GH levels as risk factors for diabetes [22]. Møller et al. have studied insulin resistance in acromegalic patients before and after adenomectomy [23]. They have demonstrated that insulin resistance in acromegalic

11 Page 11 of 27 patients is relieved after adenomectomy. Normalization of glucose response to insulin may explain why in our patients, glucose metabolism was improved and prevalence of diabetes reduced to that of the general population. The precise mechanism of insulin resistance in presence of chronic GH excess is still not clear. GH has a short term insulin-like effect but chronic exposure impairs insulin response. Cross-talk between GH receptors and insulin-receptors have been demonstrated, both sharing some post receptor signaling pathway elements. Insulin receptor substrate 1 (IRS-1) and phosphatidylinositol 3-kinase (PI3K) appear as two of the main elements of this cross-talk. To complicates matter more, in-vitro and in-vivo studies have shown different effects. A decrease in glucose transporter, has also been hypothesized as one of the mechanisms of hyperglycemia. Uncoupling between PI3K and it s downstream signals, and IRS inhibition could reduce IR response. This effect appears predominantly in the liver. Secondary hyperinsulinism may also down-regulate IR (for a review of these mechanisms, see Dominici et al., [24] and Möller et al [25]). More recently, liver IGF-1 receptor deficient mice have shown that IGF-1 plays also a role in insulin sensitivity [26]. Modulation of glucocorticoid metabolism through the activity of 11 ß-hydroxysteroid dehydrogenase may also add another level to the complexity of glucose metabolism regulation under GH and IGF-1 [27]. We looked also at the evolution of diabetes in patients with «dissociated» GH and IGF-1 (patients that normalize only IGF-1 or GH) hypothesizing a difference in glucose metabolism in these two groups. No significant difference was found when comparing these two group of patients. Another question that arises is the impact of different treatments of acromegaly in the evolution of diabetes. In our registry, there are different approaches in treatment (some centers treat each patient with somatostatin analogs for at least 3 months before surgery, others not; some centers use more frequently radiotherapy,etc.) and the size of the cohort do not allow for a statistically significant stratification of patients based on different treatment scheme. Ronchi et al. compared two somatostatin analogs (Octrotide LAR vs Lanreotide SR) describing a decrease in insulin resistance

12 Page 12 of 27 with both drugs, Octreotide LAR appearing more detrimental to diabetes than Lanreotide. [28] In a paper by Colao et al. [29], where a group of patients were treated for 5 years with somatostatin analogs, no significant change was noted in the prevalence of diabetes during the study duration although HOMA-R, and HOMA-β significantly decreased. These authors have also assessed glucose tolerance (GT) in patients treated for 12 months by SS analogs [30]. Disease control, GH levels and initial GT were the major predictors of the evolution of GT in this cohort. Mazziotti et al. studied the effect of high doses of SSA on glucose metabolism of patients whose acromegaly was inadequately controlled with standard doses of SSA showing a low impact on glucose metabolism which was related with biochemical markers of acromegaly [31]. Møller et al [23] have showed a complete reversal of insulin resistance and glucose and lipid metabolism in acromegalic patients cured by adenomectomy. In a study by Kinoshita et al. [32], glucose metabolism in patients cured by surgery was improved when patients had preserved ß-cell function. Colao et al. have also tried to compare glucose metabolism in four group of patients treated either by SSA alone, surgery alone or a combination of two, showing and improvement in some patients under SSA, and a similar deterioration in all group[33]. In order to compare the impact of different treatments on glucose metabolism, we will probably need prospective studies with strict treatment protocols. 5 Conclusions This study, based on the French Acromegaly Registry helps to further approach the prevalence of diabetes in acromegaly. In this population, the prevalence of diabetes was estimated to be 22.3%. Growth hormone and IGF-1 levels did not appear as predictive factors for the presence of diabetes. In contrast, age, BMI and hypertension were significant risk factors. These risk factors are also found in the general population of type 2 diabetics. Acromegaly could thus represent a strong sensitizing factor revealing a latent type 2 diabetes. Management of this complication is an intrinsic part of the management of acromegaly.

13 Page 13 of 27 The Authors do not have any conflict of interest. The French Acromegaly Registry was founded with the help of an unrestricted educational grant from IPSEN, Novartis and Pfizer. The members of the French Acromegaly Registry not in the author list are: Sonnet E.,CHU de Brest, France,Cazabat L., Cochin Hospital, Paris, France, Hieronimus S., Hopital de l Archet,Nice, France, Gaillard R., CHU de Lausanne, France, Enjalbert A., CHU la Timone, Marseille, France, Raingeard I. CHU de Montpellier, France, Dupuy O. Hôpital d instruction des Armées Béguin, St. Mandé, France, Beckers A. CHU de Liège Belgium. References [1] Melmed, S. Medical progress: Acromegaly. The New England Journal of Medicine 355(24), , December (2006). PMID: [2] Colao, A., Ferone, D., Marzullo, P., and Lombardi, G. Systemic complications of acromegaly: epidemiology, pathogenesis, and management. Endocrine Reviews 25(1), , February (2004). PMID:

14 Page 14 of 27 [3] Holdaway, I. M., Ra jasoorya, R. C., and Gamble, G. D. Factors influencing mortality in acromegaly. The Journal of Clinical Endocrinology and Metabolism 89(2), , February (2004). PMID: [4] Puder, J. J., Nilavar, S., Post, K. D., and Freda, P. U. Relationship between disease-related morbidity and biochemical markers of activity in patients with acromegaly. The Journal of Clinical Endocrinology and Metabolism 90(4), , April (2005). PMID: [5] Dekkers, O. M., Biermasz, N. R., Pereira, A. M., Romijn, J. A., and Vandenbroucke, J. P. Mortality in acromegaly: a metaanalysis. The Journal of Clinical Endocrinology and Metabolism 93(1), (2008). PMID: [6] Møller, N. and Jorgensen, J. O. Eects of growth hormone on glucose, lipid, and protein metabolism in human subjects. Endocr Rev 30(2), , Apr (2009). 9 [7] Jaffrain-Rea, M. L., Moroni, C., Baldelli, R., Battista, C., Maffei, P., Terzolo, M., Correra, M., Ghiggi, M. R., Ferretti, E., Angeli, A., Sicolo, N., Trischitta, V., Liuzzi, A., Cassone, R., and Tamburrano, G. Relationship between blood pressure and glucose tolerance in acromegaly. Clinical Endocrinology 54(2), , February (2001). PMID: [8] Jaffrain-Rea, M., Minniti, G., Moroni, C., Esposito, V., Ferretti, E., Santoro, A., Infusino, T., Tamburrano, G., Cantore, G., and Cassone, R. Impact of successful transsphenoidal surgery on cardiovascular risk factors in acromegaly. European Journal of Endocrinology / European Federation of Endocrine Societies 148(2), , February (2003). PMID:

15 Page 15 of 27 [9] Drange, M. R., Fram, N. R., Herman-Bonert, V., and Melmed, S. Pituitary tumor registry: a novel clinical resource. J Clin Endocrinol Metab 85(1), , Jan (2000). [10] Giustina, A., Barkan, A., Casanueva, F. F., Cavagnini, F., Frohman, L., Ho, K., Veldhuis, J., Wass, J., Werder, K. V., and Melmed, S. Criteria for cure of acromegaly: a consensus statement. The Journal of Clinical Endocrinology and Metabolism 85(2), , February (2000). PMID: [11] World health organization: Definition, diagnosis and classification of diabetes mellitus and its com- plications: Report of a WHO consultation. part 1: Diagnosis and classification of diabetes mellitus. Geneva, World Health Org. (1999). [12] Kusnik-Joinville, O., Weill, A., Salanave, B., Ricordeau, P., and Allemand, H. Prevalence and treatment of diabetes in france: trends between 2000 and Diabetes Metab 34(3), , Jun (2008). [13] Bringer, J., Fontaine, P., Detournay, B., Nachit-Ouinekh, F., Brami, G., and Eschwege, E. Prevalence of diagnosed type 2 diabetes mellitus in the french general population: the INSTANT study. Diabetes & Metabolism 35(1), (2009). [14] Daly, A. F., Rixhon, M., Adam, C., Dempegioti, A., Tichomirowa, M. A., and Beckers, A. High prevalence of pituitary adenomas: a cross-sectional study in the province of liege, belgium. J Clin Endocrinol Metab 91(12), , Dec (2006). 10 [15] Alexander, L., Appleton, D., Hall, R., Ross, W. M., and Wilkinson, R. Epidemiology of acromegaly in the newcastle region. Clin Endocrinol (Oxf ) 12(1), 71 9, Jan (1980).

16 Page 16 of 27 [16] Bengtsson, B. A., Edén, S., Ernest, I., Odén, A., and Sjögren, B. Epidemiology and long-term survival in acromegaly. a study of 166 cases diagnosed between 1955 and Acta Med Scand 223(4), (1988). [17] Nabarro, J. D. Acromegaly. Clin Endocrinol (Oxf ) 26(4), , Apr (1987). [18] Holdaway, I. M. and Ra jasoorya, C. Epidemiology of acromegaly. Pituitary 2(1), 29 41, Jun (1999). [19] Mestron, A., Webb, S. M., Astorga, R., Benito, P., Catala, M., Gaztambide, S., Gomez, J.-M., Halperin, I., Lucas-Morante, T., Moreno, B., Obiols, G., de Pablos, P., Paramo, C., Pico, A., Torres, E., Varela, C., Vazquez, J.-A., Zamora, J., Albareda, M., and Gilabert, M. Epidemiology, clinical characteristics, outcome, morbidity and mortality in acromegaly based on the Spanish acromegaly registry (registro espanol de acromegalia, rea). Eur J Endocrinol 151(4), , Oct (2004). [20] Bex, M., Abs, R., T Sjoen, G., Mockel, J., Velkeniers, B., Muermans, K., and Maiter, D. Acrobel the belgian registry on acromegaly: a survey of the real-life outcome in 418 acromegalic sub jects. Eur J Endocrinol 157(4), , Oct (2007). [21] Wass, J. A., Cudworth, A. G., Bottazzo, G. F., Woodrow, J. C., and Besser, G. M. An assessment of glucose intolerance in acromegaly and its response to medical treatment. Clin Endocrinol (Oxf ) 12(1), 53 9, Jan (1980). [22] Biering, H., Knappe, G., Gerl, H., and Lochs, H. [prevalence of diabetes in acromegaly and cushing syndrome]. Acta Med Austriaca 27(1), (2000).

17 Page 17 of 27 [23] Møller N, Schmitz, Jøorgensen JO, Astrup J, Bak JF, Christensen SE, Alberti KG, Weeke J. O. Basal- and insulin-stimulated substrate metabolism in patients with active acromegaly before and after adenomectomy. J Clin Endocrinol Metab May;74(5): [24] Dominici FP, Argentino DP, Muñoz MC, Miquet JG, Sotelo AI, Turyn D. Influence of the crosstalk between growth hormone and insulin signalling on the modulation of insulin sensitivity. Growth Horm IGF Res Oct;15(5): [25] Møller N, Jørgensen JO. Effects of growth hormone on glucose, lipid, and protein metabolism in human subjects. Endocr Rev Apr;30(2): Epub 2009 Feb 24. [26] Yakar S, Liu JL, Fernandez AM, Wu Y, Schally AV, Frystyk J, Chernausek SD, Mejia W, Le Roith D. Liver-specific igf-1 gene deletion leads to muscle insulin insensitivity. Diabetes May;50(5): [27] Neggers SJ, van der Lely AJ. Modulation of glucocorticoid metabolism by the GH-IGF-I axis. Endocr Dev. 2011;20: Epub 2010 Dec 16. [28] Ronchi, C., Epaminonda, P., Cappiello, V., Beck-Peccoz, P., and Arosio, M. Effects of two different somatostatin analogs on glucose tolerance in acromegaly. J Endocrinol Invest 25(6), 502 7, Jun (2002).

18 Page 18 of 27 [29]Colao A, Auriemma RS, Galdiero M, Lombardi G, Pivonello R. Effects of initial therapy for five years with somatostatin analogs for acromegaly on growth hormone and insulin-like growth factor-i levels, tumor shrinkage, and cardiovascular disease: a prospective study. J Clin Endocrinol Metab Oct;94(10): [30]Colao A, Auriemma RS, Savastano S, Galdiero M, Grasso LF, Lombardi G, Pivonello R.Glucose tolerance and somatostatin analog treatment in acromegaly: a 12-month study. J Clin Endocrinol Metab Aug;94(8): Epub 2009 Jun 2. [31]G. Mazziotti, T. Porcelli, F. Bogazzi, G. Bugari, S. Cannavo, A. Colao, R. Cozzi, L. De Marinis, E. degli Uberti, S. Grottoli, et al. Effects of high-dose octreotide LAR on glucose metabolism in patients with acromegaly inadequately controlled by conventional somatostatin analog therapy. Eur. J. Endocrinol., March 1, 2011; 164(3): [32] Kinoshita Y, Fujii H, Takeshita A, Taguchi M, Megumi M, Oyama K, Yamada S, Takeuchi Y. Impaired glucose metabolism in Japanese patients with acromegaly is restored after successful pituitary surgery if pancreatic {beta}-cell function is preserved. Eur J Endocrinol Feb 1. [33] Colao A, Auriemma RS, Galdiero M, Cappabianca P, Cavallo LM, Esposito F, Grasso LF, Lombardi G, Pivonello R. Impact of somatostatin analogs vs surgery on glucose metabolism in acromegaly: results of a 5-year observational, open, prospective study. J Clin Endocrinol Metab Feb;94(2): Epub 2008 Nov 11.

19 Page 19 of 27 Legend of figures: Figure 1: Density graphs comparing diabetic and non-diabetic patients age and BMI at diagnosis. Figure 2: Density graphs comparing GH and IGF-1 values of diabetic and non-diabetic patients at diagnosis. Figure 3: Prevalence of diabetes based on acromegalic patients age. Figure 4: Density graph comparing the age of patients with and without hypertension at diagnosis. Figure 5: Comparison of diabetic and non-diabetic patients age after control of acromegaly. Table 1: Continuous variables. Table2: Count variables. Table3: Prevalence of diabetes in the literature.

20 Age at diagnosis BMI at diagnosis Page 20 of 27 Density Diabetics Others Density Diabetics Others Y.O. BMI

21 Page 21 of 27 Mean GH IGF 1 Density Diabetics Others Density Diabetics Others ng/ml % of normal values

22 % of diabetics vs Age Prevalence of diabetes by age group Page 22 of 27 % of diabetics % Y.O Age

23 Page 23 of 27 Age at diagnosis Hypertension Others Density Y.O.

24 Age at last evaluation Page 24 of 27 Density Diabetics Others Y.O.

25 Page 25 of 27 Diabetics Non diabetics All patients Variable Mean Median Q1 Q3 Mean Median Q1 Q3 Mean Median Q1 Q3 Age BMI 29,5 28,7 25,98 32,65 26, ,1 27,4 26,6 24,3 29,7 Duration of acromegaly before diagnosis 9, ,25 14, , Mean GH (ng/ml) 24,6 11,45 11,3 63,8 20,6 9, ,4 10 9,5 50 IGF-1 (% , , ,1 176, ,1 194,7 371,3 PRL 51,8 10,6 6 24,62 31,2 14 7, ,6 13 7,2 26 Table1: Continuous variables

26 Page 26 of 27 Diabetics Non diabetics All patients Sex Males Females Hypertension Rhumatological Pit. deficienciecies Adenoma Mico Macro Table 2: Count variables

27 Page 27 of 27 Authors n Prevalence Maestron, 2004 [19] ,6 % Biering, 2000 [22] % Vitale, 2005 [23] ,5 % Bengtsson, 1988 [16] % French register, ,3 % Nabarro, 1987 [17] ,8 % Gordon, 1962 [24] % Ronchi, 2002 [28] % Bexx, 2007 [20] ,3 % Table 3: Prevalence of diabetes in acromegaly