The Effect of 6 Months’ Treatment With Pasireotide LAR on Glucose Metabolism in Patients With Resistant Acromegaly in Real-World Clinical Settings

Objective The aim of the study was to evaluate glucose metabolism, as measured by glycated hemoglobin (HbA1c) levels and the need for antidiabetic medical treatment, in patients with acromegaly resistant to first-generation somatostatin receptors ligands (SRLs) treated with pasireotide long-acting release (LAR) in real-world clinical practice. Biochemical control of acromegaly, as measured by growth hormone (GH) and insulin-like growth factor 1 (IGF-1) levels, was also assessed. Study Design Two-center retrospective cohort of consecutive patients with acromegaly treated with first-generation SRLs at maximum doses, who had not achieved biochemical disease control. After SRLs were discontinued, patients were given pasireotide LAR 40 mg i.m. every 28 days. The dose was increased to 60 mg i.m. in patients for whom adequate control was not achieved after 3 months. Patients were given dietary and lifestyle advice, and antihyperglycemic treatment was modified as needed. Main Outcome Measures Biochemical disease control parameters (GH and IGF-1 concentration), as well as HbA1c level at baseline and after 6 months. Results In total, 39 patients with acromegaly were enrolled. GH concentration decreased (Δme =-1.56 µg/L, range -21.38–3.62, p <0.001) during 6 months of pasireotide LAR treatment. A worsening of metabolic status was observed, with an increase of median HbA1c (Δme =0.40%, range -0.20%–2.30%, p <0.001), and antihyperglycemic treatment intensification in 23 (59.0%) patients. The median decline in IGF-1 concentration was: -283.0 µg/L, range -682.7–171.6, p <0.001. IGF-1 reached the age- and gender-specific upper level of normal in 23 (59%) patients. Conclusions Pasireotide LAR is an effective therapeutic option in patients with acromegaly refractory to first-generation SRLs. However, this therapy may result in pasireotide LAR-associated hyperglycemia, which requires early and aggressive antidiabetic medical therapy to prevent glucose homeostasis alterations.


INTRODUCTION
Acromegaly is usually caused by pituitary adenoma secreting growth hormone (GH). Excess GH accompanied by increased insulin-like growth factor 1 (IGF-1) levels results in cardiovascular and metabolic complications. As the morbidity and mortality rate in acromegaly is significantly associated with IGF-1 and GH level, the aim of medical therapy is to ensure target IGF-1 level within age-and gender-specific normal ranges and to lower and normalize GH levels (1)(2)(3)(4). Biochemical control of acromegaly assessed by a random GH level <2.5 µg/L in a recently published longitudinal study was associated with a significantly lower risk of developing diabetes mellitus as well as cardiovascular system disorders overall compared to those not controlled (5).
The treatment of choice is selective transsphenoidal adenomectomy performed by an experienced neurosurgeon. The average cure rate is 65%, but adenoma size, extrasellar extension, and cavernous sinus invasion are the most reliable factors predicting a complete tumor resection in individual patients (3,4).
First-generation somatostatin receptor ligands (SRLs) (octreotide or lanreotide) are used either as preoperative or adjuvant therapy. In patients for whom this treatment failed (acromegaly resistant to first-generation SRLs), pasireotide longacting release (LAR) is a reliable second-line therapeutic option. Pasireotide LAR is a multireceptor-targeted SRL, binding to four of five somatostatin receptor subtypes (SSTR 1,2,3,5), with the highest affinity for SSTR-5. Somatostatin receptors are also expressed in pancreatic islet cells; therefore, pasireotide LAR adversely affects glucose metabolism (6). This side effect of pasireotide LAR has been clearly identified in phase 3 clinical trials (2,7) and discussed later by other authors (8,9). However, the disease itself may destabilize glucose metabolism, as excess GH impairs insulin sensitivity and stimulates gluconeogenesis (10). Clinically, acromegaly is associated with cardiovascular and metabolic complications, and up to 50% of patients with acromegaly have prediabetes or diabetes (11,12). Therefore, it is very important to understand the impact of pasireotide LAR on glucose metabolism not only in the clinical trial environment but also in daily clinical practice.
The aim of the study was to evaluate glucose metabolism, as measured by glycated hemoglobin A1c (HbA1c) levels and the need for antidiabetic medical treatment, in patients with acromegaly resistant to first-generation SRLs treated with pasireotide LAR in real-world clinical practice. Biochemical control of acromegaly, as measured by GH and IGF-1 levels, was also assessed.

Study Design and Participants
Between January 2019 and June 2020, we enrolled all 39 patients with acromegaly treated with first-generation SRLs at maximum doses, for whom biochemical disease control had not been achieved. Patients were from two tertiary referral centers. A history of type 1 diabetes mellitus was the only exclusion criterion from the study. During the recruitment of our patients switching to pasireotide LAR was the only option of second line medical therapy in Poland as the pegvisomant or cabergoline were not reimbursed by the National Health Fund. After discontinuation of treatment with SRLs, patients were given pasireotide LAR 40 mg i.m. every 28 days. The dose was increased to 60 mg i.m. in 24 patients (61.5%) for whom adequate control was not achieved after 3 months. Patients were given dietary and lifestyle advice, and antihyperglycemic treatment was modified when needed.
The study protocol was reviewed and approved by the local Bioethics Committees.
OGTT was not performed at V0 if the diagnosis of diabetes was known earlier (prior to study entry; n=1). Additionally, at V6 OGTT was performed in all patients except for those with diabetes already diagnosed at V0 (n=4).
Additionally, patients were categorized by their acromegaly biochemical control status, defined as:

Laboratory Tests
HbA1c was measured using the ion-exchange high performance liquid chromatography (HPLC) method, certified by the National Glycohemoglobin Standardization Program (NGSP) and the International Federation of Clinical Chemistry (IFCC), and the Bio-Rad D-10 Hemoglobin Testing System ® (Bio-Rad Laboratories, CA, USA). The linear range was 18-179 mmol/mol (3.8%-18.5%) and coefficient of variation (CV) <2.5%. R a n d o m s e r u m G H w a s d e t e r m i n e d u s i n g t h e electrochemiluminescence (ECLIA) immunoassay on the Elecsys ® analyzer (Roche Diagnostics, Germany). GH was calibrated against the pure 22kDa recombinant growth hormone (WHO IRP 98/574). Functional sensitivity was 0.05 µg/L and limit of detection was 0.03 µg/L, respectively. Serum IGF-1 was measured by enzyme-linked immunosorbent assay (ELISA; LDN, Germany). Analytical sensitivity was 9.75 µg/L. IGF-1 values were compared with age-and genderstandardized normal values.

Statistical Analysis
Descriptive statistics were used to present the study population at baseline. For categorical variables, the number and percentage of occurrences were reported; for continuous variables, the following were provided: mean and standard deviation (SD), median and the range as well as the number of observations.
The dynamics of evaluated parameters were analyzed using the Wilcoxon exact test and McNemar-Bowker test. Spearman's Rho correlation coefficient (r S ) and the Mann-Whitney exact test were applied to examine the relationships between changes in HbA1c and age, baseline GH or IGF-1 levels and gender.
All statistical analysis was performed using IBM ® SPSS ® Statistics 25. A p-value below 0.05 was considered statistically significant.

Demographic and Clinical Characteristics at Baseline
In total, 39 patients with acromegaly were enrolled at two centers: 20 at one center, 19 at the other. Mean age was 47.2 ± 12.7 years, and 16 (41%) patients were female ( Table 1).
Normal glucose tolerance was observed in 15 (38.4%) patients. In total, 20 (51.3%) patients were diagnosed as prediabetic and were given dietary advice or metformin or both (dose range: 500-2,450 mg/d). Diabetes was present in four (10.3%) patients ( Figure 1). Of these, two patients were treated with metformin only (1,000-1,500 mg/d), one was treated with metformin and insulin, and one with metformin and gliclazide. We observed a borderline association between increasing age and worsening glucose metabolism status (p trend =0.053). Median HbA1c percentage at baseline was 5.60% (range 5.0%-7.5%). HbA1c correlated with age (r S =0.44; p=0.005).

Glucose Metabolism After 6 Months of Treatment
During the study period, the glucose metabolism status worsened (Figure 1, p<0.001): prediabetes progressed to diabetes in 12 of 20 (60%) patients; in 12 of 15 (80%) patients with normal glucose tolerance, prediabetes or diabetes had developed. Overall, three patients had normal glucose tolerance.   After 6 months of treatment with pasireotide LAR median HbA1c increased by D me =0.40%, (range -0.20%-2.30%, p<0.001) ( Figure 2); this change did not correlate with age (r S =0.056; p=0.736). The mean HbA1c percentage did not differ at 6 months between controlled and uncontrolled patients regardless of the criteria for control of acromegaly (IGF-1, GH, or a combination of both).
During 6 months of pasireotide LAR treatment, antihyperglycemic therapy was intensified in 23 (59%) patients (p<0.001) (Figure 3). In patients taking metformin at baseline, the median dose increased by 1,000 mg/day (range 0-2,000 mg/ day; p<0.001). The number of patients taking metformin during the study increased from 14 to 26 (Figure 4), and in 10 patients additional drugs were introduced: glucagon-like peptide-1 (GLP-1) analogue, dipeptidyl peptidase 4 (DPP-4) inhibitor or basal insulin. Details of glucose metabolism status for individual patients who required more than one antidiabetic drug are presented in Table 2.
After 6 months of treatment 24 patients (61.5%) were receiving the 60 mg of pasireotide LAR, whereas remaining 15 patients (38.5%) the dose of 40 mg. The IGF-1 level within ULN and GH < 2.5 µg/L were confirmed in 23 (59.0%) and 21 (53.8%) patients, respectively, whereas both criteria were fulfilled in 16 (41%) patients ( Table 3). If we adopted here a more stringent criterion for biochemical control of acromegaly based on GH level <1.0 µg/L (commonly used for postsurgical assessment) we would obtain a lower rate of GH normalization, namely in eight patients (20.5%). We observed no differences in the proportion of men and women who achieved biochemical control of acromegaly by the applied definitions. However, patients with biochemical control were older than those with no control ( Table 3).

DISCUSSION
Pasireotide LAR is an effective treatment option for patients with acromegaly resistant to first-generation SRLs (7,8). Our study in Polish clinical practice confirms these results. Improvement in acromegaly parameters (GH, IGF-1) was observed in all patients participating in the study. We found that patients who achieved disease biochemical control were on average older than those who did not, and this observation was consistent across the different definitions of disease control. A possible explanation could be the more aggressive course of the disease in younger patients than the older ones. A recently published retrospective analysis of 96 patients with acromegaly found similar differences; the authors suggested different acromegaly phenotypes: one in young patients who   respond poorly to medical treatment and the other in older patients who respond well to medical therapy (14). Our study confirms the increased risk of pasireotide LARassociated impairment of glucose metabolism, which is well known and an expected adverse event of such a treatment, and has been observed in previous clinical trials (2,7,15). We found that after 6 months of treatment with pasireotide LAR, the proportion of patients with normal glucose metabolism decreased from almost 40% to less than 10%. These data should be viewed in light of two facts: one that diabetes is a very common complication in patients with acromegaly, and the other that medical treatment of acromegaly may result worsening of glucose metabolism. The prevalence of diabetes in this patient population is explained by the impact of GH on insulin sensitivity and gluconeogenesis (10) as already mentioned, and has been demonstrated by large epidemiological studies; e.g. Petrossians et al. showed that the prevalence of diabetes at the time of acromegaly diagnosis was 27.5% in the Liège Acromegaly Survey (16). In our cohort, this proportion was lower: slightly over 10% of patients were diagnosed with diabetes but more than 50% had prediabetes when the treatment with pasireotide LAR started. The lower proportion of patients with overt diabetes leads us to the second fact, namely effects of treatment. Even though our patients at the time of treatment switch were not controlled with first-generation SRLs, they were likely to have long-standing disease, successfully treated at least during certain periods. According to a meta-analysis of 31 studies including 619 patients, the first-generation SRLs had no effect on FPG or HbA1c (17), and pegvisomant may even moderately improve insulin sensitivity (18,19).
As hyperglycemia and diabetes are long-term risk factors for cardiovascular disease, early and effective management of this complication is important. Furthermore, proper management of hyperglycemia increases the chance of maintaining long-term compliance with pasireotide LAR (20).
According to Samson (20), it is possible to predict which patients will develop diabetes based on FPG levels. In the Sheppard et al. study, diabetes (defined as glucose level >125 mg/dl) developed in 45% of patients with baseline FPG levels >100-125 mg/dl and in 19% of patients with baseline FPG levels <100 mg/dl (21). Similar findings were observed for baseline HbA1c level: 55% of patients with impaired glucose tolerance (HbA1c levels 5.7%-6.5%) developed diabetes, while in patients with normal glucose tolerance (HbA1c levels <5.7%) only 23% developed it (21). Other studies also show that baseline HbA1c may be the most important predictor of development of diabetes in patients treated with pasireotide LAR (22)(23)(24)(25). Thus, baseline FPG and HbA1c levels should be assessed before starting and during pasireotide LAR therapy.
In the Colao et al. study, 44.4% of patients in the pasireotide LAR group received antidiabetic medications, with metformin   and sulfonylureas the most commonly used (2). In our study, an antidiabetic drug had to be initiated in approximately one-third of patients. In addition, most patients who were treated with metformin at baseline required a dose increase and in several patients two or more antidiabetic drugs had to be administered. The mechanism of pasireotide LAR-related hyperglycemia needs to be understood to introduce optimal treatment (20). Insulin-producing pancreatic b-cells mainly express SSTR-2 and SSTR-5; pasireotide LAR suppresses insulin secretion because of its high SSTR-5 affinity. Although glucagon-producing a-cells express SSTR-2, the pasireotide LAR inhibitory effect on glucagon secretion is only modest (10).
Considering the above, it is essential to closely monitor blood glucose levels in patients treated with pasireotide LAR. Hyperglycemia should be managed with proper antidiabetic drugs. No specific recommendation has been published for acromegaly patients with hyperglycemia (20). However, there are expert statements on short-acting pasireotide-induced hyperglycemia management in Cushing's disease (2,26). According to the summary of product characteristics for pasireotide LAR, the European Medicines Agency and Food and Drug Administration recommend managing pasireotide LAR-induced hyperglycemia with antidiabetic drugs, using standard guidelines for treatment of hyperglycemia (27,28). Conforming to these guidelines, unless contraindicated, metformin is the preferred treatment for hyperglycemia (29). However, some studies show that DPP-4 inhibitors and GLP-1 receptor agonists are more effective options than metformin or sulfonylureas, and authors suggest them as the first-line treatment (20,30). The maximum daily dose of metformin is 2550 mg in the USA and 3000 mg in the European Union. Additionally, doses above 2000 mg are generally associated with little additional efficacy and, in some cases, poorer tolerability (29). Therefore, it seems reasonable to consider adding an incretin-based therapy after reaching a dose of 2000 mg of metformin in order to increase the hypoglycaemic effect and avoid accumulation of metformin side effects.
We have also found that the age of the patients positively correlated with HbA1c at baseline and that HbA1c was not affected by the acromegaly control status. This observation reflects the situation in the general population and is also consistent with the findings of the French Acromegaly Registry study (31).
We analyzed the impact of 6 months' pasireotide LAR treatment on acromegaly parameters and glucose metabolism. Some studies show that FPG and HbA1c levels rise rapidly for the first 3 months and then remain stable without further antidiabetic therapy (7,21,23,25). Further research is necessary, therefore, to confirm this in clinical practice.
Our study was conducted in clinical practice in Poland where, for years, only the first-generation SRLs (octreotide and lanreotide) were reimbursed for treatment of acromegaly (32,33). GH-receptor antagonist pegvisomant, which is particularly useful in patients with associated diabetes was not financed from public funds at the time the study began. Therefore, when the pasireotide LAR treatment program started, patients generally presented with more advanced disease and with a relatively high prevalence of complications of acromegaly. In other words, their metabolic status and complications were greater than expected immediately after diagnosis of acromegaly resistant to firstgeneration SRLs. This is the probable explanation of why our cohort of patients was particularly prone to deterioration of their metabolic status. The other limitation of our study lies in its retrospective observational nature. This type of study allows generation of real-world evidence and informs the medical community about the effects of treatment in an unselected patient cohort; however, it does not allow for exclusion of bias and does not ensure robustness and completeness of data. For this reason, we decided to assess glucose metabolism with HbA1c and not FPG.
We confirm that treatment with pasireotide LAR may result in carbohydrate metabolism disorders, which in most patients requires intensification of antidiabetic therapy. Therefore, all patients treated with pasireotide LAR should be given dietary and lifestyle advice and have glucose-related parameters closely monitored. These findings do not undermine our other observation that pasireotide LAR is an effective therapeutic option in patients with acromegaly refractory to firstgeneration SRLs. Additionally, if properly managed, it is also a safe therapeutic option.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The studies were conducted according to Good Clinical Practice guidelines and the Declaration of Helsinki. The studies involving human participants were reviewed and approved by Bioethics