More than a decade of real-world experience of pegvisomant for acromegaly: ACROSTUDY

Objective To report the final long-term safety and efficacy analyses of patients with acromegaly treated with pegvisomant from the ACROSTUDY. Design Global (15 countries), multicentre, non-interventional study (2004–2017). Methods The complete ACROSTUDY cohort comprised patients with acromegaly, who were being treated with pegvisomant (PEGV) prior to the study or at enrolment. The main endpoints were long-term safety (comorbidities, adverse events (AEs), pituitary tumour volumes, liver tests) and efficacy (IGF1 changes). Results Patients (n = 2221) were treated with PEGV for a median of 9.3 years (range, 0–20.8 years) and followed up for a median of 7.4 years (range, 0–13.9 years). Before PEGV, 96.3% had received other acromegaly treatments (surgery/radiotherapy/medications). Before PEGV treatment, 87.2% of patients reported comorbidities. During ACROSTUDY, 5567 AEs were reported in 56.5% of patients and of these 613 were considered treatment-related (in 16.5% of patients) and led to drug withdrawal in 1.3%. Pituitary imaging showed a tumour size increase in 7.1% of patients; the majority (71.1%) reported no changes. Abnormal AST or ALT liver tests occurred in 3.2% of patients. IGF1 normalization rate improved over time, increasing from 11.4% at PEGV start to 53.7% at year 1, and reaching 75.4% at year 10 with the use of ≥30 mg PEGV/day in an increasing proportion of patients. Conclusion This comprehensive review of the complete cohort in ACROSTUDY confirmed the overall favourable benefit-to-risk profile and high efficacy of PEGV as mono- and combination therapy in patients with an aggressive course/uncontrolled/active acromegaly requiring long-term medical therapy for control.


Introduction the drug increases its biological half-life. PEGV is administered by subcutaneous injections and
is mainly used as a monotherapy, but can also be used in combination with SRLs or DAs. The main biochemical marker to monitor the efficacy of PEGV is serum IGF-1 (4,13).
In the initial pivotal trials, PEGV was generally well tolerated in patients with acromegaly treated for up to 18 months, and serum IGF-1 normalization was achieved in 97% of patients with ≥12 months of daily PEGV (14, 15). PEGV was approved in Europe in 2002 (16) and in the US in 2003 (17) for patients with acromegaly that could not be adequately controlled by surgery and/or radiation therapy and/or medication. The drug was later approved in the US for use as a primary medical therapy when surgery and/or radiation therapy fail to control the disease or when these therapies are not appropriate (17).
The global, multicentre, non-interventional ACROSTUDY was established in 2004 as a post-authorization safety surveillance (PASS) requested by the European Medicines Agency (EMA) to evaluate the long-term safety and efficacy outcomes of PEGV as prescribed in routine clinical practice for acromegaly. Previous reports revealed reassuring safety outcomes in monotherapy and combination therapy (18,19,20,21,22,23,24,25,26). With ACROSTUDY concluded in December 2017, this report presents the final safety and efficacy outcomes in the complete cohort of 2221 patients with up to 14 years of follow up.

Study design
Patients were enrolled in ACROSTUDY on an ongoing basis and followed for a minimum of 5 years. The PASS commitment of evaluating the safety of 5-year treatment with PEGV in at least 1000 patients with acromegaly was fulfilled in 2013 (19,20). The study was extended as a voluntary PASS (voluntary extension) to study safety (especially changes in glucose) and quality of life measures, which allowed approximately 400 patients already enrolled in the ACROSTUDY to continue (rollover patients), and was open to an additional 100 new patients with acromegaly who were either treatment naïve or semi-naïve (no PEGV treatment within 6 months prior to enrolment) at study entry. The study took place in 15 countries where PEGV was authorized for acromegaly treatment. Treatment doses and schedules were at the discretion of the treating investigators. Patients were assessed during routine clinical practice and followed up until December 2017, when ACROSTUDY was terminated.
The ACROSTUDY data reported here were collected in compliance with, and consistent with, the most recent version of the Declaration of Helsinki. In addition, the study adhered to all applicable local laws and regulatory requirements in the countries involved. Patients with acromegaly who were being treated or just starting PEGV treatment prior to enrolling in ACROSTUDY were included. Some European study sites were able to enrol paediatric patients (<18 years). Pituitary imaging within 6 months prior to study enrolment was recommended for all patients. During the voluntary extension, only adult patients (≥18 years) were eligible for enrolment. Patients were excluded from enrolment if they had participated in any other investigational trial for acromegaly in the previous 6 months, required surgery to decompress the tumour or non-medical therapy due to visual field loss, cranial nerve palsies or intracranial hypertension, or had allergies to PEGV or its ingredients. Women who were pregnant or lactating were also not enrolled. In the voluntary PASS, re-enrolment of patients who were discontinued from the ACROSTUDY was not allowed. Informed consent was obtained from all patients.

Safety
The baseline was defined as the start of PEGV treatment. Safety data included adverse events (AEs), liver tests and pituitary tumour imaging. All reported AEs, serious AEs (SAEs), AEs of special interest (e.g., administration site reactions, hepatobiliary-related AEs, and changes in tumour size), and deaths were assessed. All AEs were coded using the Medical Dictionary for Regulatory Activities dictionary (MedDRA version 20.1). Concomitant medication and acromegaly-related comorbidities were collected from treating physicians using a standardized questionnaire. Any disorders with onset after enrolment were reported as AEs. Safety endpoints related to liver tests included the percentage of patients with elevated transaminases (≥3-fold of upper limit of normal [ULN]) in the alanine aminotransferase (ALT) or aspartate aminotransferase (AST) tests by visit window, and the percentage of patients with any liver test elevation (alkaline phosphatase, bilirubin, gamma-glutamyltransferase; ≥ULN). Pituitary tumour images were collected and read locally at baseline, and at months 6 and 12 post treatment, and then annually as determined by the investigator. A significant change in tumour size was defined as a >3 mm change in the largest diameter of a pituitary microadenoma or a >20% change in tumour volume of a macroadenoma (>10 mm). If significant changes were determined by local radiologists, all available images for that patient were requested to be re-assessed by a central reader (this was not required during the voluntary extension). When available, assessments of pituitary tumour volume (increased, decreased, or unchanged) relative to baseline were summarized by visit window.

Efficacy
Serum IGF-1 levels were measured at local and central laboratories. Efficacy data analysed included the proportion of patients by IGF-1 values (<lower limit of normal [LLN], normal, >ULN according to the laboratory reference values) and proportion of patients who achieved and maintained IGF-1 level within the normal range. An "IGF-1 controlled" status was assigned to those with a value in the normal range from their latest IGF-1 measurement. An IGF-1 status of "not controlled" was assigned to those with an IGF-1 value higher than ULN or lower than LLN.
Fasting blood glucose and HbA1c parameters were obtained through routine clinical practice and percentages of patients with values outside of normal range were summarized.

Statistics
No pre-specified statistical hypotheses were tested in the study. All data were summarized with descriptive statistics. Percentages of patients experiencing an AE, liver test elevation, change in tumour size, or other laboratory value outside of the normal range were assessed for a specified time period or time point. IGF-1 levels were categorized according to the laboratory reference values (normal, >ULN, or <LLN) and summarized by years of PEGV treatment with mean daily dose included. Comorbidities were quantified as reported by the treating physicians.

Demographics
A total of 2221 patients from 14 European countries and the US ( Table 1) participated in ACROSTUDY and were included in the safety population. Of these, 434 were rollover patients and 110 were naïve/semi-naïve patients, which were enrolled as part as the voluntary PASS extension. The majority of patients were Caucasian (92.4%) and the percentage of male to female was similar. Overall, patients had a median age of 41.1 years at acromegaly diagnosis, 49.7 years at the start of PEGV treatment, and 51.5 years at study enrolment. The maximum duration of PEGV treatment was 20.8 years, with a median of 9.3 years. Patients were followed in ACROSTUDY for up to 13.9 years (median 7.4 years). A total of 11 paediatric patients were enrolled (0.5%); 5 were between 2 and 11 years old and 6 were 12 to <18 years old. Pituitary function data at baseline showed deficiencies for follicle-stimulating hormone/luteinizing hormone (FSH/LH) in 37.8% of patients, thyroid-stimulating hormone (TSH) in 28.9%, and adrenocorticotropic hormone (ACTH) in 28.2% ( Table 2).

Acromegaly treatment
Prior to PEGV treatment, almost half (48.1%) of patients had been treated with both medical and surgical therapies, 21.6% had all three interventions (surgery, radiotherapy, and medication) and 18.8% medication alone ( Table 1). Before PEGV start, 65.7% of patients received SRLs only, 31.2% received SRLs combined with DAs and 3.0% used DAs alone ( Fig. 1). At PEGV start, the most commonly prescribed treatment was PEGV monotherapy (55.1%), followed by PEGV combined with an SRL (34.3%; Fig. 1). Treatment pattern slightly changed over time, with the use of PEGV monotherapy ranging between 47.0% to 55.1%, while use of PEGV/SRL decreased from 34.3% to 24.4% over the years.
AEs of special interest related to administration-site condition were reported in 3.5% of patients, with most of them related to PEGV treatment (3.2%; Table 4). Most common treatment-related AEs were lipohypertrophy (reported under skin and subcutaneous tissue disorders), for 1.2% (n = 27) of patients and injection-site reaction AEs, for 0.8% (n = 18) of patients. Administration-site condition AEs led to PEGV withdrawal (or dose reduction) in 1.1% of patients, with lipodystrophy/lipohypertrophy being the cause 13 times.

Pituitary tumour imaging
The percentage of patients receiving a pituitary MRI scan was 40.3% at year 1 and 45.2% at year 2, which decreased over time to 31.0% at year 5 and 4.4% at year 15. A total of 1795 patients had ≥1 pituitary imaging result after PEGV initiation (Fig. 3  which led to PEGV withdrawal in 1.7% of patients. No liver failure was reported in the study. Two patients had liver enzyme abnormalities that satisfied potential Hy's Law criteria (ALT/AST >3x ULN and peak total bilirubin result >2x ULN). However, both had other conditions that were believed by the investigators to have led to their elevated liver test results.

IGF-1 normalization
At PEGV start, 11.4% of the patients (n = 1546) had an IGF-1 concentration within the normal reference range, while 88.4% had an IGF-1 >ULN. The percentage of patients with IGF-1 levels within the normal range increased from 53.7% at year 1 to 63.3% at year 5, and remained above 60% (63.3%-79.3%) throughout most of the study (Fig. 4). Overall, 62.3% of patients with IGF-1 data available beyond baseline assessment achieved an IGF-1 normalization at the last observation. The degree of IGF-1 normalization was accompanied by an increase in PEGV doses; thus, mean daily doses in patients with IGF-1 normalization increased from 14.0 mg at year 1 to 18.2 mg at year 10 ( Fig. 5). PEGV doses in patients with active acromegaly also increased over time and were higher than those in controlled patients at most years.

Discussion
The present final report summarizes data on the complete cohort of 2221 patients who participated in ACROSTUDY with a maximum follow-up time of 13.9 years. Overall, PEGV appeared to be well tolerated over a median duration of 9.3 years of use. The main finding of this study was that PEGV had a favourable safety profile in clinical practice, especially concerning pituitary tumour volumes and liver tests. However, it should be noted that most subjects were Caucasian, and the sample size of non-Caucasian subjects was too small to determine a risk-tobenefit profile. We also note that the pharmacokinetic and pharmacodynamic properties of PEGV are not significantly different between Asian and Western patients so we would expect similar results (27, 28,29,30).
The prescribing information of PEGV (17) and Endocrine Society guidelines recommend monitoring tumour size and liver tests during the course of PEGV treatment (1). and as such, pituitary tumour growth and elevated liver enzymes were the main safety concerns with PEGV use. In the full cohort, central MRI reading showed tumour size increase in 3.7% (alone or in combination with a decrease), similar to those reported in the German Pegvisomant Observational Study (3.1%) (31) and other reports of ACROSTUDY (2.2%-3.2%) (19, 22, 25).
This low incidence suggests that PEGV does not promote tumour volume increases. The observed events could indeed represent the result of tumour recurrence/regrowth due to SRL withdrawal or could simply reflect the natural history of an aggressive tumour (32,33).
Interestingly, the incidence of pituitary tumour volume increase based on local MRI readings was higher than that based on central reading (7.2% vs 3.0%), showing the possible difference in assessment by local analysis as noted previously (19). In smaller, real-life studies of PEGV, a higher incidence of tumour growth was reported (6.5%-9.4%; central reading not mentioned), close to that assessed by local reading in ACROSTUDY (28,29,30,34). Over time, the percentage of patients with pituitary MRI scans decreased, likely reflecting perceived patient status and standard of care.
We found here a low incidence of liver enzyme elevation during PEGV treatment, not different from previous reports (18,19,20,22,26). Only 3.2% of patients with normal baseline ALT/AST value had an elevated ALT/AST (>3x of ULN) at any time point during the follow up.
Overall, increases in liver enzymes were transient in most patients and liver failure was not reported. Elevated liver tests in 5.2%-9.3% of patients were observed in other real-world studies of PEGV (29, 30, 31). It is noteworthy that 30%-40% of patients in ACROSTUDY received combined PEGV/SRLs (with or without other drugs), which could be expected to result in higher rates of transaminase elevation compared with PEGV alone (23,35,36). However, possibilities of transient elevations occurring between PEGV initiation and ACROSTUDY start or between scheduled liver tests cannot not be excluded (23). In addition, the majority of the hepatobiliaryrelated AEs were considered to have causes other than PEGV and only 1.7% of patients discontinued PEGV due to these events.
PEGV therapy may cause lipodystrophy, a disorder of adipose tissue, at the injection site (37,38). In ACROSTUDY, most of the administration-site condition AEs (reported for 2.0% of patients) were considered related to PEGV. As lipodystrophy may be associated with escape (loss of biochemical control in patients previously controlled) from PEGV (38), the PEGV injection site should be monitored for lipodystrophy and frequent injection-site changes (25).
We observed that IGF-1 normalization rate progressively improved with PEGV treatment over time: more than half (53.7%) of patients within 1 year and by a maximum of 79.3% at later years. Almost two-third of patients with IGF-1 data after baseline (62.3%) had documented IGF-1 normalization at last follow-up, consistent with previous findings of 63%-73% in the two interim ACROSTUDY reports published in 2012 (19) and 2018 (20). While the IGF-1 normalization rate was also lower here than in the pivotal clinical trials leading to PEGV approval (14, 15), our findings agree with other real-life studies including the German Pegvisomant Observational, retrospective Brazilian, retrospective Argentinian, and Japanese post-marketing surveillance studies (28,29,30,31). Any discrepancy could be attributed to lack of uniform dose titration, inadequate patient compliance, and different IGF-1 assays or lack of titration/normalization criteria used in these real-world settings versus clinical trials, which have more strict trial criteria (18,20,31). For example, a higher rate of IGF-1 normalization (85%-90%) was observed in two studies of PEGV-treated patients at tertiary care hospitals, where tight dose titrations and close follow-up were more likely (34,39). In addition, patient selection may be skewed to those who had more aggressive acromegaly as most patients in ACROSTUDY were enrolled at European sites, where PEGV is indicated as a second-line medication; here almost half of patients had failed surgery and other medications and another 20% also underwent radiation (19, 22). Finally, with the improved biochemical control over time, use of PEGV doses higher than 30 mg daily also increased; however, mean doses were similar between patients with controlled and uncontrolled acromegaly.
In this complete ACROSTUDY analysis of full cohort, the use of PEGV monotherapy remained constant over time (55.1% at baseline to 53.3% at year 15), while use of PEGV in combination with SRL and/or DA decreased (44.9% to 31.0%). This differs slightly from a previous interim analysis of ACROSTUDY, which reported a higher proportion of patients receiving combination therapy over time (20% in 2003 vs 54% in 2012) (23). Differences in these analyses include the number of samples used and how rates were measured, but these final results may reflect the normal clinical course of action in acromegaly treatment. Reasons for the decrease of PEGV in combination with SRL/DA over time are likely multifactorial and varied due to treatment practices across countries and centres, but could be due to toleration issues, patient preference, cost of the combination therapy, radiation, and optimization of biochemical control. Use of combination treatment in acromegaly patients may benefit patients with aggressive acromegaly (6), particularly those partially resistant to first generation SRLs and with large/invasive tumours (41), similar to those treated with PEGV and pasireotide (42, 43). A recent study showed that low-dose SRL plus weekly PEGV represents a potential novel dosing option for achieving cost-effective, optimal biochemical control in patients with uncontrolled acromegaly requiring combination therapy (44).
PEGV effectiveness may vary depending on patient characteristics. An observational retrospective study found that PEGV resistance was associated with higher BMI and was more frequent with BMI >30 kg/m 2 (45). Lower baseline of GH, IGF-1 and IGF-1 x ULN were associated with disease control, which was more frequent with baseline IGF-1 <2.7x ULN (45). Therefore, higher starting PEGV doses and a more rapid up-titration may be necessary in obese patients and in those with IGF-1 levels >2.7x ULN (45). The recent Pituitary Society Guidelines suggest that patients with diabetes mellitus and those with a higher BMI require higher doses of PEGV and more rapid up-titration to achieve IGF-1 normalization (46).

An earlier, 4-year, longitudinal interim analysis of ACROSTUDY (n = 1762) explored
the effects of PEGV on glucose metabolism in patients with (n = 109) or without diabetes (47).
In patients with diabetes mean blood glucose decreased by 20.2 mg/dl from baseline to year 4, while mean HbA1c remained unchanged (47). At year 1, IGF-1 normalization rate was slightly lower in patients with diabetes than those without (52.1% vs 57.4%) (47). Overall, the mean daily PEGV dose was higher in patients with diabetes than without (18.2 mg/day vs 15.3 mg/day) (47). These results were similar to those reported in an 18-study meta-analysis of interventional studies (48), which showed that PEGV alone or combined with SRLs improved glucose metabolism.
Due to the nature of long-term observational, non-interventional studies, ACROSTUDY Patients with acromegaly have significant morbidity and increased mortality if not appropriately treated (2). This study evaluated associated-acromegaly complications in a large cohort in a real-world clinical setting and more than 90% of patients had more than one comorbidity at study entry, with hypertension, diabetes and osteoarthritis being the most frequent ones. Not surprisingly, many patients had colon and thyroid tumours, which highlights the need for screening for these conditions. This large final cohort from ACROSTUDY followed a high number of unselected broad-range of patients from multiple countries for almost 14 years of follow up, allowing for a better understanding of dose use, efficacy and the safety profile of PEGV and treatment pattern changes over time.

Conclusions
The global, non-interventional ACROSTUDY provides safety and efficacy data of PEGV on the        3 a n = 1126; b n = 1081; c n = 2207; d included race as "Oriental".
1  1 Table 4 Adverse events and deaths in Full Analysis Population (n = 2221).