Recurrence in acromegaly: two tertiary centers experience and review of the literature

Background Recurrence of acromegaly after successful surgery is a rare event, but no clear data are reported in the literature about its recurrence rates. This study aimed to evaluate the recurrence rate in a series of acromegalic patients treated by transsphenoidal surgery (TSS) with a long follow-up. Methods We retrospectively analyzed data from 283 acromegalic patients who underwent TSS at two pituitary units in Milan (Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico and IRCCS Humanitas Research Hospital). The diagnosis and recurrence of acromegaly were defined by both elevated IGF-1 levels and a lack of GH suppression based on appropriate criteria for the assay used at the time of diagnosis. Results After surgery, 143 patients (50%) were defined as not cured, 132 (47%) as cured and 8 (3%) as partially cured because of normalization of only one parameter, either IGF1 or GH. In the cured group, at the last follow-up (median time 86.8 months after surgery), only 1 patient (0.7%) showed full recurrence (IGF-1 + 5.61 SDS, GH nadir 1.27 µg/l), while 4 patients (3%) showed only increased IGF1. In the partially cured group at the last follow-up, 2/8 (25%) patients showed active acromegaly (IGF-1 SDS + 2.75 and + 3.62; GH nadir 0.6 and 0.5 µg/l, respectively). Conclusions In the literature, recurrence rates range widely, from 0 to 18%. In our series, recurrence occurred in 3.7% of patients, and in fewer than 1%, recurrence occurred with elevation of both IGF-1 and the GH nadir. More frequently (25%), recurrence came in the form of incomplete normalization of either IGF-1 or GH after surgery.


Introduction
Acromegaly is a chronic and rare disease characterized by GH hypersecretion, most frequently caused by GH-secreting pituitary tumors arising from somatotroph cells [1].Rare causes of acromegaly are pituitary carcinoma and ectopic production of GHRH by both central hypothalamic tumors and peripheral neuroendocrine tumors [2].Active acromegaly causes a twofold increase in mortality, and malignancies are the leading cause of death, followed by cardiovascular disease [3,4].Biochemical control of the disease has been confirmed to reduce the excess mortality in the general population [2,4].
The Endocrine Society guidelines recommend transsphenoidal (TSS) surgery as first-line therapy in most patients [1].The outcomes of the procedure are dependent on various factors, such as the tumor size, degree of invasion and experience of the surgical team [5]; with experienced pituitary surgeons, TSS surgery results in an initial remission rate ranging from 40 to 85%, depending on the series and the size of the tumor [1].
The criteria to define both diagnosis and remission of acromegaly have changed over the years, mainly due to the availability of more sensitive GH assays [6].As for remission, the current criteria were established by the Consensus Conference in 2010 and confirmed by the 2014 Clinical Practice Guidelines as IGF-1 levels in the age-and sex-adjusted normal range and, in patients with elevated or equivocal serum IGF-1 levels, random GH levels < 1.0 mcg/l using commercially available immunoassays 12 weeks after surgery [1,7].When random GH levels are detectable, current guidelines suggest measuring the GH nadir after a glucose load (OGTT), with levels greater than 0.4 mcg/liter in an ultrasensitive chemiluminescent assay (CLIA) defining active disease [1,7].
In the past, when GH levels were measured with classical radioimmunoassay (RIA) [8], random GH levels < 2.5 mcg/l and nadir GH levels after a glucose load < 2 mcg/L were accepted as standards for the remission of disease [6].In 1999, the Consensus Conference in Cortina defined cure criteria as IGF-1 levels in the age-and sex-adjusted normal range and a random GH value < 2.5 mcg/l or a GH nadir after OGTT < 1 mcg/l when immunoradiometric assays (IRMA) became available [8][9][10].These past assays were competitive and were relatively insensitive, typically exhibiting a low detection limit [8].
It is critical to differentiate patients with persistent disease from those in remission to start with the correct therapy and follow-up.To do so, the timing of biochemical testing after surgery is very important because IGF-1 has a longer halflife and therefore takes more time to decrease and normalize than GH does [5].
Recurrence is defined as a relapse into GH hypersecretion after initial postsurgical remission, as defined by the criteria for cure [11].In the literature, the reported recurrence rates are variable, and studies reporting recurrence rates are heterogeneous in terms of population, criteria for the definition of cure, and timing of biochemical testing after surgery.Many of these papers report surgical series in which recurrence was not the main aim of the study (Table 1).
To date, there are no clear clinical suggestions about the timing of long-term follow-up or the criteria for discontinuation.In our tertiary centers in Milan, most acromegalic patients, even when cured by surgery, stay under lifelong follow-up, which brings personal and social costs.
This study aimed to evaluate acromegaly patients successfully treated by TSS with a long follow-up history to establish the recurrence rate and the need for further follow-up.

Methods and study design
We retrospectively analyzed the data of 283 consecutive acromegalic patients (168 F, mean age 44.2 ± 12.9) who underwent TSS for a GH-secreting pituitary tumor between 1980 and 2020 in two specialized centers in the city of Milan: Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico (245 patients) and IRCCS Humanitas Research Hospital (38 patients), who underwent at least one year of follow-up after surgery.
Hormonal data included IGF-1 and GH Nadir.The GH nadir was defined as the lowest GH value at any time during a 2-h oral OGTT.For this, blood samples were collected at 30, 60, 90, and 120 min after glucose load.
Study eligibility criteria included age greater than 18 years, a diagnosis of acromegaly based on the combination of both an elevated serum IGF-1 level and lack of GH suppression based on criteria appropriate for the assay used at the time of diagnosis: GH > 2 µg/L by RIA, GH > 1 µg/L by modern IRMA or GH > 0.4 µg/L by CLIA.They also had to have MRI evidence of pituitary adenoma and to have undergone transsphenoidal endonasal surgery (TSS).
IGF-1 values are expressed as standard deviation scores (SDSs).We obtained SDS values according to the methods provided by Chanson and colleagues [12].
The local ethics committee (Comitato Etico Milano Area 2) approved this protocol study.

Study design
All patients had a biochemical diagnosis of acromegaly according to the criteria used at the time of evaluation.
All patients underwent thinly sliced magnetic resonance imaging (MRI) or computerized tomography (CT) of the sellar region to evaluate the presence of a pituitary tumor.Tumors were classified by their maximum diameter as microadenomas (≤ 10 mm) or macroadenomas (> 10 mm).
All patients underwent TSS at one of the two neurosurgery units after the confirmation of the acromegaly diagnosis.34 patients received medical treatment before   We analyzed the data of all patients at the first follow-up after surgery, at least 12 weeks after the procedure.We defined patient outcomes as follows: -cured when both normal serum IGF-1 levels and normal GH suppression after OGTT (defined by the criteria for the assay used at the time) were observed; -partially cured when only IGF1 level or GH suppression was normalized; -not cured in the case of both elevation of serum IGF-1 level and lack of GH suppression (defined by the criteria for the assay used at the time).
The cured and partially cured patients were evaluated at least annually for a median follow-up time of 86.8 months (IQR 31-172.9months).Their status at the last available follow-up was analyzed here.Recurrence was defined by the presence of elevated IGF-1 levels in addition to a lack of GH suppression (defined by the criteria for the assay used at the time), and the occurrence data were analyzed at the time of recurrence and at the last available follow-up.

Statistical analysis
Quantitative variables are expressed as mean ± SD when normally distributed, while quantitative variables that were not normally distributed are expressed as median (interquartile range (IQR)).Qualitative variables are expressed as number and percentages (%).
Comparisons between two groups of data were analyzed by the paired or unpaired t-tests, when variables were normally distributed.Comparisons between categorical variables were assessed by the chi-square test, or by Fisher's exact test when the expected values were < 5. Data were analyzed using GraphPad Prism (version 5.0, La Jolla, CA, USA) or SPSS (PASW Version 19.0, SPSS Inc., Chicago, IL, USA).P values (p) < 0.05 were considered statistically significant.

Results
All 283 patients had preoperative confirmation of acromegaly (mean IGF1: + 13.7 ± 9.2 SDS, median GH nadir 7.6 IQR 2.8-17.9µg/L).MRI confirmed the presence of a pituitary tumor in all patients (192 macro/91 micro), all patients underwent TSS, and all patients had pathological confirmation of a pituitary tumor positive for GH on immunohistochemical staining.

First follow-up after surgery
At the first follow-up after surgery (median distance of 3 months IQR 2-3 months), we defined 143 patients (50%) as not cured, 132 patients (47%) as cured, and 8 patients (3%) as partially cured, of whom 6 patients had normal IGF1 values but no GH suppression after OGTT and 2 patients had elevated IGF1 levels and normal GH suppression.

Cured patients at last follow-up
The median follow-up available for the total population of the study was 104.9 months (IQR 45.6-212.1 months).Within the group of cured patients, at the last available follow-up (median follow-up time 86.8 months IQR 31-172.9months after surgery), 5 out of 132 (3.7%) showed abnormalities of hormone levels.
In particular, 1 patient (0.7%) showed a biochemical status of active acromegaly with both elevated IGF1 levels (IGF1 + 5.61 SDS) and lack of GH suppression after OGTT (GH nadir 1.27 µg/L) 15 months after surgery.Both evaluations of IGF1 and GH levels, after surgery and at the time of recurrence, were analyzed in the same laboratory with the same assay (CLIA).
Four other patients (3%) showed an increase in IGF1 levels during follow-up despite consistently normal GH suppression.After multiple confirmations of isolated IGF1 levels above the upper normal limit (mean IGF1 SDS + 3.65), with a mean follow-up of 81 months from surgery, these patients started medical therapy for acromegaly.
All patients with biochemical partial or complete recurrence underwent a new MRI of the pituitary region.In all cases, this demonstrated neuroradiological stability with no signs of tumor recurrence or regrowth (Table 2).At the time of diagnosis and before surgery, the only patient with true recurrence had a microadenoma, while of the other 4 patients 2 had a microadenoma (50%) and 2 had a macroadenoma (50%).

Partially cured at last follow-up
In the partially cured group, 2 (25%) patients showed, 12 and 37 months after surgery, a biochemical status of active acromegaly (IGF1 SDS + 2.75 and + 3.62; GH nadir 0.6 and 0.5 µg/l, both assessed with CLIA; all data are summarized in Table 2).In this group of patients, no signs of tumor regrowth were found on pituitary MRI (Table 2).Both recurrent patients started medical therapy for acromegaly after evidence of both elevated IGF1 and GH nadir levels.
Interestingly, one of the patients who experienced recurrence had normal IGF1 levels but a lack of GH suppression at the first follow-up after surgery, while the second patient showed normal GH suppression but elevated IGF1 levels.
Both patients had a macroadenoma at the time of diagnosis before surgery.
The percentage of recurrent patients in the partially cured group was significantly higher than that in the cured group (p = 0.008).We did not find any other significant difference between the recurrent group and the nonrecurrent group, which is likely due to the small number of patients in the sample.We also analyzed the possible role of different neurosurgeons, however, the two centers did not significantly differ in recurrence rates: in the first center of 111 patents cured or partially cured after surgery 5 had a recurrence (4.5%); in the second of the 21 patients cured or partially cured after surgery 2 had a recurrence (9.5%); p = 0.3.Our data suggest that an incomplete cure after surgery (considering both GH or IGF1 levels) is associated with a higher likelihood of recurrence.
Moreover, we did not find any significant difference in biochemical status after surgery or pathological findings between patients with and without complete recurrence in the partially cured group, again probably because of the small number of patients (IGF1 after surgery + 1.6 ± 0.6 SDS vs. + 1.3 ± 0.8 SDS, GH random 3.4 ± 0.3 vs. 2.7 ± 0.7, Ki67% was less than 1% in all patients).

Discussion and conclusions
Recurrence rates in acromegalic patients successfully treated by surgery range widely in the literature.Different series were published with a percentage of recurrence that ranges from 0% [13] to 18% [15] (Table 1).
Many reasons could underlie these important differences: the series vary in terms of numbers and population characteristics, timing of follow-up and, most importantly, the criteria used to define active acromegaly (all data are shown in Table 1).Many of the series are surgical ones, whose aim was to evaluate the surgical remission rather than recurrence after surgery and thus frequently there is little precision or attention to hormonal data and recurrence features.Other possible reasons for such a wide range may be the tumor dimension, invasion, and aggressiveness [32] or different genetic causes of acromegaly.In addition, when defining the correct recurrence rate, the timing of follow-up is crucial, especially the timing of the first follow-up after surgery.Guidelines suggest that the first follow-up should be at least 12 weeks after surgery in order to correctly categorize the outcome of surgery [1].This timing takes into consideration the fact that the decline in IGF-1 levels is more delayed compared with the decline in GH values, likely due to the differential half-life of IGF-binding proteins [33].
An earlier follow-up after surgery may not reflect the real biochemical status and may be confusing for both patients and practitioners.Moreover, in some series, the definition of recurrence was based on either IGF1 levels alone, random GH or GH suppression after OGTT (with different cutoffs), or the simultaneous presence of more than one of these parameters.As expected, stricter criteria are linked to a lower prevalence of recurrence than criteria that do not consider both IGF1 and GH suppression.For example, the paper by Biermasz and colleagues [15], which reports the highest percentage of recurrence (18%), analyzed a small cohort of patients, of whom 75% had MRI evidence of macroadenomas and 15% had evidence of invasive tumors.The timing of the first follow-up after surgery was not specified; it was generically defined as early postoperative, and only one parameter, either GH or IGF1, defined recurrence.A small population with aggressive tumors, with no strict diagnostic criteria and a possible misleading biochemical definition after surgery, will likely be linked to a higher percentage of recurrence (data shown in Table 1) [15].
In contrast, we found less than 1% true recurrence (0.7%), in agreement with the largest series published to date on this topic, by Nomikos and colleagues, who reported a recurrence rate of 0.4% in a cohort of 506 acromegalic patients with strict diagnostic criteria applied three months after surgery [13].
These very low recurrence rates are also in agreement with other previous studies using similar methodological approaches in terms of the timing of follow-up and diagnostic criteria, thus confirming that acromegaly recurrence is very rare in patients successfully treated with surgery [5,13,17].
A particular situation in our cohort is represented by the 3% of patients who, after multiple confirmations of elevated IGF1 levels, started a medical therapy for their partial recurrence despite the persistence of normal GH suppression.During medical therapy, we could not repeat testing for GH suppression, so we are not able to exclude the possibility that these patients were evaluated during an early phase of relapse and that over time they would have shown complete biochemical recurrence.On this topic, a recent and interesting long-term prospective study by Freda and colleagues demonstrated that by performing repeatedly OGTT in a small series of acromegalic patients defined as cured after surgery, a few of them had an elevation in the GH nadir levels after OGTT over time (using a cutoff relative to the 97.5th percentile of nadir GH in 100 healthy subjects, which was ≤ 0.14 µg/l with DSL IRMA), to the point of even reaching a biochemical definition of complete recurrence (GH nadir > 0.4 µg/l).Interestingly, recurrence was more likely in patients who had a GH nadir > 0.14 µg/l on initial postsurgical testing [14].This study further underlines the importance of performing OGTT in patients after surgery to correctly interpret the hormonal status and plan the follow-up.Unfortunately, the retrospective nature of our study could not better investigate these aspects.
Moreover, in our retrospective cohort, the small number of recurrences did not allow us to investigate, with statistically significant results, the presence of predictive factors.Our study reports an important recurrence rate (25%) in patients with incomplete biochemical remission after surgery, with a statistically significant difference compared to the group of patients completely cured after surgery (p = 0.008).Once more, as the number of patients with incomplete remission was small, we could not define whether values of GH or IGF1 were more important in the postoperative evaluation.Nevertheless, our data underline the need to focus on incompletely cured patients.
Overall, these findings, as also suggested by Freda and colleagues, support the role of a correct and complete postsurgical follow-up, suggesting the need to perform an OGTT in all patients, as opposed to guidelines that recommend performing OGTT only in patients with basal random GH levels > 1 µg/l [14].
GH measurement and its suppressibility by OGTT are not perfect measures, since current diagnostic GH nadir values should probably be better individualized according to age, sex, BMI and estrogenic status [34,35].This could be a factor that limited our ability to identify patients with altered GH secretion after surgery and at the time of recurrence.
Finally, our study suggests that recurrence, when present, is limited to the biochemical status.Interestingly, none of our patients showed either tumor regrowth or neuroradiological modification.This might have important practical clinical implications because, if confirmed by further larger studies, it could indicate the possibility of planning a biochemical follow-up alone without the need for radiological imaging, reducing the patient's burden and costs.
In conclusion, our data confirm that acromegaly recurrence is very rare in patients successfully treated after surgery, even after a long follow-up time.More frequently (25%, p = 0.008), recurrence occurs in the form of incomplete normalization of either IGF1 or GH after surgery.As also suggested by previous studies using our approach, the right timing and the use of strict biochemical criteria at first follow-up after surgery are crucial to correctly define the surgical outcome and plan an appropriate follow-up.Indeed, our data suggest that most acromegalic patients with complete remission after surgery may be considered definitely cured, with no need for strict follow-up.In contrast, patients with partial remission after surgery need strict follow-up.

Limitations
We acknowledge that this study is limited by the retrospective nature of the analysis.Moreover, our data were collected throughout a follow-up period of almost 40 years, during which time laboratory assays and biochemical criteria have changed several times reflecting technical improvements.Finally, the relatively small number of recurrences and partial remissions limits the strength of our conclusions and precludes a definitive identification of prognostic factors and clinical suggestions.

Table 1 (
* BDI: Biochemical discrepancy type I (normal IGF1 levels + random serum GH ≥ 1.0 μg/L); BDII: Biochemical discrepancy type II (high IGF-I + random serum GH < 1.0 μg/L) ° IGF-I levels up to 1,25-fold the ULN determined 3 months after TNS in acromegaly patients was associated with long-term remission, regardless of GH levels, even in patients initially considered as surgical failure.IGF-I ≤ 1,25-fold the ULN had 100% sensitivity and 96% specificity to predict long-term remission # Recurrence rate was calculated on 61 patients initially in remission with available data 15 years after surgery (

Table 2
Individual hormonal and radiological data of patients, both cured and partially cured at first follow-up, who showed recurrence of disease after surgery