In this work, by means of the Delphi methodology, a panel of Italian endocrinologists assessed factors determining fgSRL resistance, second-line medical treatment [18,19,20], and T2DM’s role in patient management. IGF-1 is unequivocally established as the primary parameter for monitoring treatment efficacy in clinical practice. The panel revealed low confidence and interest in the use of cabergoline in second-line, while all the other choices are considered according to patient’s condition. Hyperglycemia is an important issue, and it could influence the choice of therapy, especially in certain categories of patients.

Incidence, definition of resistance to therapy with fgSRLs, and predictors of response

Although both GH and IGF-1 levels are central in defining resistance to treatment (statement 1.4), the panel acknowledges that IGF-1 is the “gold standard” (statement 1.1). The role of IGF-1 is consolidated [21, 22], as it was similarly recognized as determinant also in 2017. On the other hand, the agreement on the role of GH results weaker compared to 2017 (75% vs 83% in 2023 vs 2017, respectively), in line with recent consensus [3].

In case of discrepant IGF-1/GH values, resistance is primarily determined by the levels of IGF-1(statement 2.4). Indeed, in one third of acromegalic patients, GH levels may give discrepant information compared to IGF-1, likely due to assay or cut-off issues, GHR polymorphism, timing of post-surgical assessment, or different biological significance of the two parameters [11, 23,24,25]. However, no consensus was reached on some statements on the definition of acromegaly under control when IGF-1 and GH levels show different responses to therapies (statements 2.5 and 2.6). In fact, for some of the experts, normalization of GH should be a clinically relevant goal irrespective of IGF-1 normalization, possibly because of the fear of a prospective increase of IGF-1 levels in the absence of GH control; however, this concept has not been included in new Consensus recommendations [3]. Nevertheless, the responses suggest potentially significant implications for clinical practice, since it could involve treatment intensification or increased follow-up even in cases of IGF-1 normalization. This might also indicate a preference for a different second-line drug that lowers GH rather than IGF-1 in cases of discrepancies.

Due to IGF-1 assay variability, particularly in mildly elevated cases, experts stress the importance of repeated evaluations over time before confirming resistance (statement 1.2). This aligns with the 2017 Delphi, albeit with lower agreement (91% in 2017 and 83% in 2023). No consensus was reached regarding whether slightly elevated IGF-1 levels should be considered an indication of effective treatment (statement 1.3). Indeed, not all the guidelines suggest as acceptable target levels up to 1.3 × ULN [4, 11]. A less ambitious target in patients who underwent radiotherapy is acceptable since progressive improvement of disease activity’s control is expected (statement 2.2). However, IGF-1 levels slightly above the normal range are not considered an appropriate target in elderly patients with a recent diagnosis, where the disease could be less aggressive (statement 2.1). Despite acceptable range in these two conditions changed between the two surveys (up to 1.5 x ULN in 2017 and up to 1.3 x ULN in 2023), this did not affect the consensus, corroborating the idea of a possible stricter IGF-1 target nowadays.

Overall, biochemical and clinical parameters should be more frequently checked in uncontrolled patients (statement 2.3) and should be integrated into the definition of resistance to fgSRLs (statement 3.1 and 3.3). In particular, a comprehensive evaluation of acromegaly, possibly through AcroDAT ® [26] and SAGIT® [27, 28], is considered valuable (statement 3.4). The absent or modest (< 20%) adenoma shrinkage is less relevant than biochemical measures in defining resistance (statement 3.2; agreement reached only in the second round); this may be due to the lack of evidence on a specific diameter/volume threshold to define clinically relevant shrinkage, particularly in small post-surgical remnants where MRI evaluation may be impacted by different artifacts [24]. Side-effects, such as hyperglycemia, should not be included in the definition of resistance to fgSRLs (statement 3.5), likely due to the reported marginal impact of fgSRLs on glucose homeostasis [29].

The possibility to predict resistance to fgSRLs would allow personalized interventions in patients identified at risk, avoiding long and ineffective attempts to control the disease [7, 11, 30]. Familiar history of pituitary adenomas and positive genetics are recognized as a predictive parameter for development of resistance (statement 4.1), particularly in young patients [31]. Absence of SSTRs and expression of molecular markers of aggressiveness in surgical pituitary tissue are strongly identified as markers of expected poor response to fgSRLs (statement 4.4–4.5). This is of great interest, since it implies that experts rely substantially on novel pathological techniques, recently found not to be routinely available also in Pituitary Tumor Centers of Excellence [32,33,34].

Finally, size of adenoma is not believed to be a predictor of resistance to fgSRLs (statement 4.2), while no consensus was reached on the role of elevated post-surgical GH/IGF-1 (statement 4.3); this is an interesting finding since virtually all guidelines report on the low likelihood of biochemical control in these two conditions. In the experience of part of the group, therefore, control could still be obtained in this context with fgSRLs. This opinion may be based on the long diagnostic delay in acromegaly, which is one of the main determinants of adenoma overgrowth and excessive GH and IGF-1 levels, with no implications on potential response to treatment [35, 36].

Biochemical and clinical determinants for the choice of second-line therapy

In case of ineffective control of acromegaly with fgSRLs, combination therapy with cabergoline [11] did not meet the consensus as further therapeutic step for the panelists (statement 5.1), even if, recently, this combination showed a good IGF-1 normalization rate (30–58% of cases) [37]. Different factors may have contributed to the absence of consensus, including limited controlled studies, off-label use of cabergoline, availability of alternative therapeutic options, weak GH suppression by this drug and potential long-term detrimental effects (of heart valves especially in the context of acromegalic cardiopathy). Cabergoline is not perceived as valuable option also when looking at compliance (statement 10.1), a finding supported by a recent real-world analysis of US administrative claims data, which reported lowest adherence and persistence for this drug [30].

In line with guidelines, experts agree that PEGV monotherapy (statement 5.2), in combination with fgSRLs (statement 5.3), or PASI (statement 5.4) could be used in patients resistant to fgSRLs [5, 11]. Although PEGV in initial studies was only used as monotherapy [38, 39], ACROSTUDY real-world cohort showed that PEGV is used in up to 50% of cases in combination with fgSRLs [40], albeit with differences among countries [41].

Guidelines suggest that PEGV monotherapy is useful in patients with glucose metabolism disorders and/or without problems related to the pituitary mass [11]. The Delphi panel agreed that the presence of T2DM motivates the choice of PEGV as monotherapy (statement 6.3), with increased consensus compared to 2017 (86% vs 72%), indicating a better knowledge of this positive impact of PEGV [12, 42]. In resistant patients, the switch to PEGV monotherapy is motivated by slightly elevated or frankly elevated IGF-1 levels (statement 6.1 and 6.2), with a stronger agreement achieved in this second cohort of patients compared to patients with a lower level of elevation of IGF-1 (69% vs 86%).

Despite reported only in a minority of cases in clinical trials [43], the growth of pituitary mass upon fgSRLs discontinuation is a concern for switching to PEGV monotherapy (statement 7.1), albeit with a decreasing consensus compared to 2017 (69% vs 74%, respectively). Previous radiotherapy is a determinant for switching to PEGV monotherapy (statement 6.5), indicating that in this cohort of patients the concern for pituitary mass growth is reduced. This statement did not achieve a consensus in 2017, possibly indicating a growing reliability of new radiotherapy techniques in controlling tumor mass [44].

PEGV daily injections schedule (statement 7.3), and the possible reappearance of headache after SRLs discontinuation (statement 7.4) are also reasons for concern; however, concern on compliance achieved a borderline consensus, lower than in 2017 (67% vs 83% in 2023 vs 2017, respectively), and in a subsequent statement (10.3) panelists agreed on PEGV monotherapy (even if not daily) being a possible therapeutic strategy considering the patient's compliance and adherence (83%).

Despite being a drawback in 2017, therapy costs do not seem to be a detrimental factor for PEGV therapy anymore (statement 7.2, 9.3), likely in light of the significant efficacy demonstrated [45,46,47] on various targets including glycemia, presumably associated with a reduction of indirect costs. In fact, according to an Incremental Cost Effectiveness Ratio analysis in the Spanish NHS [48], PEGV is the most cost-effective alternative in the treatment of acromegaly in fgSRLs-resistant patients.

According to the panel, the transition to PASI should not necessarily take place after initial treatment with PEGV (either in monotherapy or combination, statement 8.1). This is in line with the 2018 consensus [11] but not with the most recent 2020 consensus, which positioned PEGV monotherapy as a second-line option, while fgSRLs + PEGV or PASI were reported as a third-line option [5]. Still, this is consistent with the European and Italian label for the drugs, which position them both as options after ineffectiveness of fgSRLs, and it is possibly related to some recent real-life data reporting an increased efficacy of PASI in fgSRLs-resistant patients [49] compared to initial studies [50].

The experts agree" on the use of PASI without particular concern in case of no medical history of T2DM (statement 8.3). Indeed, PASI could induce a worsening of the metabolic picture by inhibiting insulin and incretin secretion, especially in patients starting therapy with increased basal glycemia. No consensus was reached on the optimal treatment approach for the management of PASI-associated hyperglycemia (statement 8.4 and 8.5). Still, a multicenter, randomized, open-label, Phase IV study reported that, in some acromegalic patients treated with PASI, hyperglycemia could be effectively controlled by metformin, eventually followed by incretin-based therapy [51].

Real-life studies have also identified a discrete action of PASI on headache [52, 53], as acknowledged by the experts (statement 8.2). As per PEGV, costs are not a determinant for switching to PASI monotherapy (statement 9.4), while the panel show some cost-related concern for PASI + PEGV combination (statements 9.5). This is possibly related to the significant higher costs of this combination, associated with the panel’s lower expertise on this approach.

To summarize, the panelists show a clear preference within the Italian NHS for an independent decision-making process, encompassing nearly all individual or combined therapy schedules (statements 9.1–9.4). PASI monotherapy achieves the highest consensus for compliance and adherence compared to the other second-line medical treatment options (statement 10.2), while a consensus was not achieved on PASI + PEGV combination therapy (statement 10.5).

Role of blood glucose levels in therapeutic management

Impaired glucose tolerance and T2DM are common in acromegaly [54] and guidelines suggest that T2DM should influence the choice of medical therapy. PASI is not recommended in uncontrolled T2DM patients because of the high risk of further glycemic control deterioration [4]; on the other hand, many data showed that PEGV treatment improves glucose metabolism [12, 43] and should be considered in patients with partial or no response to fgSRLs for whom glycemic control is challenging.

In the experts’ opinion, T2DM/metabolic alterations should be approached and treated as in the general population (statement 13.4, 13.5). Hyperglycemia onset should not modify acromegaly therapy, even if induced by it (statement 12.1). An HbA1c target below 7.0%, or further tailored to the patient’s age (statement 12.2, 12.4) and the management of cardiovascular risk factors (statement 12.6) are goals for acromegaly and hyperglycemia treatments. In the second round, the panel reached a consensus on a lower HbA1c target (below 6.5%, statement 12.3), a parameter that sparked disagreement in 2017. This finding, in line with the most recent T2DM guidelines [55], is possibly due to more efficacious and safer treatment options for diabetic patients, while a more stringent diabetes control is probably also seen as a proactive measure to mitigate cardiovascular mortality.

No consensus was reached on the referral of patients to a Diabetology Centre (statement 12.5), likely reflecting the different organization of Italian Endocrinology practices, where only some centers have Diabetology Units. It could be assumed that some neuroendocrinologists personally manage T2DM to avoid delocalization of patients, while others, considering acromegaly diabetes as non-specific form of diabetes, prefer referring these patients to diabetologists.

In fgSRL resistant patients, the choice of second-line depends on metabolic compensation (statement 14.1), but a consensus could not be reached on whether diabetes mellitus per se should influence the second-line medical therapy (statement 14.2) and on the addition of PEGV in these patients (statement 14.3). On the other hand, panelists do not perceive that in these cases the only possibility to start PASI is in association with PEGV (statement 14.4), probably also considering that PASI’s detrimental effect on glucose metabolism is only partially counteracted by PEGV [56]. Experts did not agree on IFG influencing the choice of second-line medical therapy, contraindicating, at least in young patients, the use of PASI (statement 11.1), or indicating the necessity to switch to PEGV monotherapy (statement 11.4). This is an interesting finding, since data from literature show that patients with IFG are at higher risk of developing hyperglycemia when switching to PASI [57], while age seems to be protective against development of T2DM while on this therapy.

FgSRLs are not considered responsible for metabolic worsening, or at least not such as to disregard their effectiveness on biochemical or tumor mass control. The onset of T2DM in well-controlled patients is not a valid reason for fgSRLs therapy discontinuation (statement 13.1), which should be accompanied by the start of an antidiabetic drug (statement 13.2). This despite evidence demonstrating significant positive impact on glucose metabolism of PEGV, regardless of IGF-1 levels [43]. Therefore, in the experts’ opinion the attainment of effective control over acromegaly is the priority. The positive effects on GH/IGF-1 secretion and tumor mass are deemed more crucial, with any potential deterioration in glycemic control being considered a separate concern. It is noteworthy that these assertions achieved a more robust consensus than in 2017. However, when T2DM treatment is ineffective, the experts agree on a switch to PEGV monotherapy also in well-controlled patients (statement 13.3); this statement did not achieve a consensus in 2017, indicating that, at least in this subpopulation, the panelists agreed that the positive effect of PEGV on glucose metabolism may be exploited to manage both conditions.

Experts agree that in a patient with history of GDM and a GH-secreting macroadenoma resistant to fgSRLs, surgical intervention should be prioritized (statement 15.4). Another option is the combination of fgSRLs and PEGV (considering the diabetogenic and tumor growth risks if fgSRLs are discontinued and PEGV is used as monotherapy) (statement 15.2). An alternative that gains consensus in the second round is shifting to PASI (prioritizing the tumor mass issue over the diabetes concern, statement 15.1).

Limitations

This work presents the limitations of the Delphi studies: a decline in response rate between the rounds of 17% was observed, but this is within the limits described for Delphi studies, especially when dealing with a large number of statements [58, 59]. The consistency of answers in subsequent interactions indicated that the statements were correctly defined. Although some of the invited experts did not participate, those taking part well represented the real-world management of acromegaly in Italy. The lack of consensus for some statements underlines the open questions that need further research to be addressed.