Clinical and Genetic Characterization of Pituitary Gigantism: An International Collaborative Study in 208 Patients

Abstract

Context:

Despite being a classical growth disorder, pituitary gigantism has not been studied previously in a standardized way.

Objective:

To characterize a large series of pituitary gigantism patients.

Design:

Retrospective, multicenter, international study.

Setting:

Tertiary referral hospital outpatients departments

Patients:

208 patients (163 males;78.4%) with growth hormone excess and current/previous abnormal growth velocity for age or final height >2SD above country normal means. 143 patients consented to genetic analyses.

Main Outcome Measures:

Growth, hormonal and symptomatic measures at baseline and during management

Results:

Median onset of rapid growth was 13.0 years and occurred significantly earlier in females than in males; pituitary adenomas were diagnosed earlier in females than males (15.8 vs. 21.5 years, respectively). Adenomas were ≥10 mm (i.e. macroadenomas) in 84%, of which extrasellar extension occurred in 77% and invasion in 54%. GH/IGF-1 control was achieved in 39% during long-term follow-up. Final height was greater in those with younger age of onset, with larger tumors and higher GH levels. Later disease control was associated with a greater difference from mid-parental height (r=0.23, P=0.02). AIP mutations occurred in 29%; microduplication at Xq26.3 -X-linked acro-gigantism (X-LAG)- occurred in two FIPA kindreds and ten sporadic patients. Tumor size was not different in X-LAG, AIP mutated and genetically-negative patients. AIP-mutated and X-LAG patients had significantly younger age at onset and diagnosis, but disease control was worse in genetically-negative cases.

Conclusions:

Pituitary gigantism patients are characterized by male predominance and large tumors that are difficult to control. Treatment delay increases final height and symptom burden. AIP mutations and X-LAG explain many cases, but no genetic etiology is seen in >50% of cases.

Introduction

Fascination with extremes of height stretches back into antiquity and the occurrence of people with gigantism contributed to popular myths across many cultures. Even with the recognition that many extreme cases of tall stature are disease manifestations, popular interest in gigantism has remained largely focused on the visual spectacle, even today. Many forms of short stature are known and screening activities in childhood are geared to identifying these cases for early investigation and intervention (1). Overgrowth is less well understood and, particularly if unaccompanied by other syndromic features like developmental delay, its diagnosis may be slower. Societal factors may contribute to this, as tall stature is seen a less “undesirable” physical feature than short stature (2). However, increased height also carries risks in terms of disease (3), and excessive final adult height carries with it distinct disadvantages, particularly skeletal and orthopedic problems (4,5). Surprisingly, the functional and psychological impacts of extreme tall stature have yet to be studied in detail.

Growth and stature are determined by highly complex processes involving genetic and environmental factors, such as, endocrine function, nutrition, vitamin status and psychosocial wellbeing (6–8). Diseases causing tall stature must be differentiated from other normal variations in height, such as constitutional tall stature in which underlying abnormalities are absent. Pathological tall stature can be isolated or syndromic, the latter usually is due to a chromosomal or genetic cause, such as Klinefelter syndrome, Marfan syndrome, Sotos syndrome among others (9). Disorders of the growth hormone axis can lead to abnormal height, the most classical of which is pituitary gigantism, usually due to over-secretion of growth hormone (GH) by a pituitary adenoma occurring before epiphyseal closure (10,11). In recent years a variety of genetic factors that predispose to somatotrope adenomas or hyperplasia have been identified. Mutations in genes such as GNAS and PRKAR1A and particularly AIP are associated with acromegaly and gigantism (12,13). X-linked acro-gigantism (X-LAG) syndrome is associated with a microduplication on chromosome Xq26.3 and leads to pituitary hyperplasia and adenomas in children and early onset gigantism beginning usually in the first year of life (14,15).

Due to its rarity and despite the recent emphasis on pathophysiological causes, the clinical presentation, evolution, complications and responses to treatment of patients with pituitary gigantism have not been studied in a large cohort. To address these issues, we conducted an international collaborative study of the features of patients with pituitary gigantism.

Methods

This was a retrospective cohort study including patients with gigantism due to a pituitary adenoma or hyperplasia. The study was performed between 2011–2013 at the Department of Endocrinology, Centre Hospitalier Universitaire de Liège, Belgium in collaboration with 46 international centers in Argentina, Australia, Belgium, Brazil, Bulgaria, Canada, Denmark, India, Italy, Finland, France, Germany, New Zealand, Romania, Russia, Spain, The Netherlands, and USA. This study was approved by the Ethics Committee of CHU de Liège (Belgian clinical trials number: B707201111968). Patients provided informed consent in their local language for genetic studies.

Eligibility criteria

The diagnosis of pituitary gigantism was defined as current or previous evidence of abnormal, progressive and excessively rapid growth velocity for age (>97th percentile, which corresponds to >+2SD), or a final height >+2SD above the mean for relevant population, associated with elevated GH/IGF-1 and imaging evidence of a pituitary lesion. Details on height sources for the countries are listed in supplemental materials.

Patient disposition

Patient information (demographics, medical and familial history, genetics, clinical examination, laboratory investigations, radiology, disease status during follow-up, treatment modalities and response to therapy) were systematically collected in each study center, recorded in the case report form (see Supplemental Material) and transmitted anonymized to the coordinating center. Overall 229 patients were enrolled; 21 cases were ineligible and were excluded (Klinefelter’s syndrome (n=5), constitutional tall stature (n=3), Sotos syndrome (n=2), obesity (n=2), ectopic GHRH secretion (n=1), and tall stature of unknown etiology without GH axis excess (n=8)). The final study population consisted of 208 patients diagnosed with pituitary gigantism (Figure S1).

Statistics

Statistical analysis (statistical computing and graphics) was performed using STATISTICA, version 10 (StatSoft), and R package, version 2.15.1 (R Core Team). Absolute numbers and percentages were used to describe qualitative and categorical data. Continuous data did not fit parameterized distributions, therefore were represented as Medians and inter quartile ranges (IQR) and non-parametric statistical tests were used for the analysis (Spearman R and Chi-square (χ2) tests for the association between variables and Mann-Whitney U and Kruskal-Wallis tests for comparison of independent subgroups). A P-value of <0.05 was designated as the level of statistical significance.

Results

Characteristics at diagnosis (Table 1)

Table 1.

Clinical characteristics of patients with pituitary gigantism.

	Total group	Males (n=163) 78%	Females (n=45) 22%
Age at rapid growth onset (year)*	13 [9;15]	13 [10;15]	11 [3;14]
Growing patients at last follow-up (%)	15.9%	13.5%	25.6%
Age of attaining final height (year)	20 [18;22]	20 [18;22]	18.5 [16;23]
Age at first symptoms (year)*	14 [10;16]	14 [11;16]	12.5 [2;14]
Age at diagnosis of PA (year)*	21 [15.5;27]	21.5 [17;28]	15.8 [10;23]
- aged ≤19 years (%)*	42.4%	37.3%	61.9%
Delay from symptoms to diagnosis (year)*	5.3 [2.0;11.0]	6.2 [3;12]	2.5 [1.0;6.0]
Z-score (SD)	3.1 [2.5;4.0]	3.1 [2.5;4]	3.1 [2.6;4.1]
Age at height measurement (year)	29 [21;37]	29 [22;38]	28 [17;35]
Difference from MPH
- absolute (cm)	20 [15;24]	19.5 [15;23]	21 [15;26.3]
- % from MPH	11.6 [8.5;14.8]	11.2 [8.4;13.7]	12.9 [8.7;16.4]
Maximal dimension of PA (mm)	22 [14;34]	21 [14;32]	24.5 [15;36]
Macroadenoma (%)	84.3%	85.1%	81.6%
-giant adenoma (%)	15%	14.6%	16.7%
Extension (%)	77.2%	77%	78%
Invasion (%)	54.5%	54.9%	53.3%
GH level at diagnosis (ng/ml)*	35.5 [14;83]	29 [12.3;64]	62.3 [27.8;95]
IGF-1 level at diagnosis (% UNL)	254.5 [189.5;359.5]	250 [188;358.5]	268 [198;421]
Prolactin co-secretion (%)	34%	31%	46.9%
Multimodal treatment (%)	32.2%	32.1%	32.5%
GH/IGF-1control at last follow-up (%)	45%	49%	32.5%
- age when GH/IGF-1 control achieved (year)	23 [17;30.5]	23.5 [17.8;31]	17.7 [11;29.5]
- GH/IGF-1 controlled ≤19 years (%)	36.6%	32.1%	55%
- GH/IGF-1 controlled before final height (%)	20.8%	19.2%	26.8%
Long-term control (%)	39%	42%	30%

^*P-value less than 0.05 when compared females and males

All continuous data presented as Median and IQR

PA= pituitary adenoma

MPH= mid-parental height

Multimodal treatment: ≥ 3 separate modalities

The pituitary gigantism population consisted of 208 patients, the majority of whom were male (n=163; 78%). The median height Z-score was +3.1 [2.5;4.0]; the median age at height measurement was 29 [21.0;37.0] years and the majority of patients (84.1%) had reached their final height (at a median age of 20 [18;22] years). The median age at the onset of rapid growth was 13 [9.0;15.0] years overall, and was significantly younger in females than in males (11.0 [3.0;14.0] vs. 13.0 [10.0;15.0] years, respectively; P=0.003). There was a median delay of 5.3 [2.0;11.0] years between first symptom onset and pituitary adenoma diagnosis; this delay was significantly shorter in females than males (2.5 [1.0;6.0] vs. 6.2 [3.0;12.0] years, respectively; P=0.03). Overall, 42.4% of patients were aged ≤19 years at diagnosis and significantly more female patients fell in this age group (P=0.004).

The most frequent first clinical sign was increased growth (approx. 75% of patients), followed by acral enlargement and facial changes (37%), headache (23%) and visual field defects (12%). Pubertal delay occurred in approximately 29% of males and females.

Nine cases of pituitary apoplexy were reported at baseline and seven patients had diffuse pituitary hyperplasia (radiological or surgical); the remainder had pituitary adenomas. Pituitary tumors were predominantly macroadenomas (84.3%), with 15% of those being “giant” adenomas. Extrasellar extension and invasion occurred in most cases (89% and 64%, of macroadenomas, respectively). Radiological characteristics did not differ between males and females.

Despite of the relatively young age of the patients, acromegalic features were present in almost all males (92%) and females (94%) at presentation (Table S2). The median shoe sizes at diagnosis were 15.0 [13.0;17.0] in males and 11.5 [10.5;14.5] in females (EU sizing− 48 [46.0;50.0] and 42 [41.0;45.0], in males and females, respectively). Among 156 cases that had cardiac assessments reported, cardiac disease had already been diagnosed in 36.5% at baseline, mainly left ventricular hypertrophy (21%) and diastolic dysfunction (10%).

Females exhibited higher median GH levels at diagnosis than males (62.3 [27.8;95.0] vs. 29 [12.3;64.0] ng/ml, respectively; P=0.009), but IGF-1 levels were similar between genders. Co-secretion of prolactin occurred in 34% overall and was more frequent in patients with invasive macroadenomas with extrasellar extension. At diagnosis, 25% of patients had a deficit in ≥1 axis; among those aged ≤19, hypopituitarism at baseline was seen in 18.3%.

Treatment and follow-up

Treatment regimens differed among centers due to the availability of medical therapies (Figure 1). Initial surgery in 177 patients was associated with control in 15%. Among 40 cases that were then re-operated, 7.5% of those were controlled. Postoperative somatostatin analogs (SSA) were used 66.7% (n=118) of patients and disease control was achieved in 34% (n=40) of them. A further 26% (n=54) of patients received primary SSA treatment, but only 7% (n=4) of these were controlled. Pegvisomant was used pre-operatively either alone (n=1) or in combination with SSA or dopamine agonists (DA; n=8); control was achieved in 4 cases. Pegvisomant was administered after surgery with SSA and/or DA in 28 patients; control was achieved in 53.5% (n=15) of these cases. A total of 63 patients were irradiated (2 had primary radiotherapy) with control in 43% (median follow-up: 168 months [62;235]); 56.5% of these also had received an SSA during follow-up. The median number of treatment modalities was 2 [1;3]. Overall, disease control was achieved in 45.6% of patients. The median duration of follow-up post treatment was 7 years [3;17] and in those followed up for ≥12 months, disease control was achieved in 39.5% of cases. There was a significant correlation between larger tumor diameter and a greater number of treatment modalities (r=0.18, P=0.02). Macroadenomas required significantly more treatment modalities than microadenomas (≥3 modalities in 50% vs. 19%, respectively; P=0.009). Median maximal tumor diameter was smaller in patients that were controlled (19 [11;25] vs. 27 mm [17.0;37.5] in uncontrolled cases; P=0.0003). However there was better control at last follow-up in those patients with tumors diagnosed aged ≤19 years than older patients (58.5% vs. 36.4%, respectively; P=0.02). Maximal tumor diameter at diagnosis was correlated with GH (but not IGF-1) levels (r=0.34, P=0.002) at diagnosis.

Figure 1.

Schematic representation of treatments used in the management of patients with pituitary gigantism. Numbers in parentheses indicate the number of patients. RadioThx =Radiotherapy, DA =Dopamine agonists, SSA = Somatostatin analogs, PegV = Pegvisomant.

*2 patients had primary radiotherapy (1 controlled); 4 recently diagnosed patients awaiting treatment; **Follow up- 168 months [62;235]

During follow-up pituitary apoplexy occurred in nine patients. Hypopituitarism rose from 25% at baseline to 64% at last follow-up. In those aged ≤19 years at diagnosis hypopituitarism was present at diagnosis in 18.3% and in 66% after treatment, with deficiency of 3 pituitary axes in 29%, and panhypopituitarism in 3%. Presence of hypopituitarism at last follow-up was significantly related to larger tumor size (30 cm [20;39] vs. 15.5 cm [10;25], P=0.006), but not to duration or control of the disease.

Seven patients (3.4%) died during follow-up; causes of death were thrombosis/embolism (n=2), hemorrhage, myocardial infarction, tumor progression, accident and suicide (n=1 each).

Growth responses

The height of each patient expressed in Z-scores above the mean and their age at last measurement are shown in Figure 2. The median height Z-score at diagnosis was higher in those who were still growing than those that had attained their final height (+4.1 SD [2.8;5.7] and +2.9 SD [2.5;3.8], respectively, P=0.004).

Figure 2.

Age and Z-scores for height of patients at last follow-up. Of the total population of 208 patients, 57 patients had an absolute height >200 cm at the last measurement (eight are still growing) and the tallest patient was 247 cm. Eleven that were controlled before the end of the liner growth had a height at the last follow-up of <+2SD (seven of these later had disease recurrence).

Excess GH/IGF-1 secretion was controlled before the end of linear growth in 20.8% of the total group; in 11 of these cases, hormonal control led to a normalization of the growth pattern and a height at last follow up that was <+2SD (Figure 2). Hormonal control at ≤19 years of age was associated with an earlier halting of linear growth than in those controlled after that age (P=0.0052). Overall, patients’ final height exceeded their MPH by a median difference of 20 cm [15;24] and no gender differences were seen.

Height Z-score and the difference from MPH correlated significantly with tumor size (r=0.2, P=0.03) and GH (but not IGF-1) at diagnosis (r=0.29, P=0.0001). Height Z-scores were also significantly greater in those that were younger both at first symptoms (r=−0.3, P=0.01) and at the start of rapid growth (r=−0.19, P= 0.01). The difference of final height from MPH depended on age when first control was achieved (r=0.23, P=0.02), being significantly lower in those with disease control aged ≤19 years than thereafter (10.9% [7.7;13.8] vs 12.7% [9.32;16.3], respectively, P=0.044). Median excess over MPH was greater in patients with hypogonadism or pubertal delay than in those with normal gonadal status (12.8% [8.8;16.3] vs. 10.7% [8.5;13.3], respectively; P=0.04).

Genetic studies

In the study population, 143 pituitary gigantism patients consented to genetic testing (AIP, MEN1, PRKAR1A, GNAS1, Xq26.3 duplication) and 46% had genetic causes or inherited syndromes (Figure 3). In total, 29% of patients were positive for AIP mutations. There were 28 FIPA patients (23 males) of whom 18 had AIP mutations. Four members of two FIPA families had Xq26.3 microduplications and X-LAG syndrome, as did a further 10 sporadic cases. In addition, seven McCune-Albright syndrome, two familial Carney Complex and one MEN1 gigantism case were observed. Fifty-four percent of the patients had no genetic cause identified.

Figure 3.

Genetic results in the study population. Numbers for each sector show the number of patients in the subgroup and its prevalence in the total group. AIP+ = AIP mutation affected; MAS = McCune-Albright Syndrome; X-LAG = X-linked acrogigantism syndrome; MEN1 = Multiple endocrine neoplasia type 1.

As compared with the AIP mutation-positive patients, those with no detected genetic cause were significantly more likely to be female, were older at first symptoms and at diagnosis (fewer cases were aged ≤19 years) and had a longer disease latency, with higher GH/IGF-1 levels, more frequent multimodal therapy and poorer overall control rates (Figure 4; Table 2). X-LAG cases were predominantly female, significantly younger at onset, but had similar tumor size and lower rates of invasion and extension as AIP mutation-positive cases (Figure 4).

Figure 4.

Comparisons of characteristics among genetically distinct groups of pituitary gigantism patients (genetically negative, AIP mutation positive (AIP+) and X-linked acrogigantism syndrome (X-LAG)), showing statistically distinct patterns of age at first symptoms (A), age at diagnosis (B) and no inter-group difference in terms of maximal tumor diameter at diagnosis (C). Panel D demonstrates the female and male predominance of X-LAG and AIP+ related gigantism cases; the genetically-negative group was also male-predominant although less markedly so than the AIPmut group.

Table 2.

Comparisons of characteristics between X-linked acro-gigantism (X-LAG) syndrome group, AIP mutation positive (AIPmut) patients and genetically negative patients.

	Group I X-LAG syndrome n=14	Group II AIPmut n=42	Group III Genetically negative n=77	P1 (I vs. II)	P2 (I vs. III)	P3 (II vs. III)
Gender (% of males)	29%	95%	78%	0.002	0.01	0.01
FIPA (number of patients/ number of families)	4 / 2	18 / 16	5 / 5
Age at onset of rapid growth (year)	1.5 [1;2]	13 [9;15]	14 [11;15]	<0.0001	<0.0001	0.09
Growing patients (%)	76.9%	12.5%	12.2%	0.005	0.0001	0.8
Age of reaching final height (year)	23.5 [18;29]	19 [18;21]	20 [18;24]	0.6	0.6	0.16
Z-score (SD)	+4.82 [+2.8;+6.1]	+3.01 [+2.5;+4.1]	+2.99 [+2.4;+4.0]	0.009	0.007	0.9
Age at height measurement (year)	5 [3;7]	26.5 [19;33.5]	29 [24.8;36.5]	<0.0001	<0.0001	0.11
Difference from MPH
- absolute (cm)	10.9 [9.4;12.4]	18.9 [16.2;26]	21.5 [14;24.5]	0.06	0.1	0.63
- relative (%)	6.52 [5.5;7.5]	10.9 [8.9;15.4]	12.7 [8.3;16.2]	0.06	0.1	0.5
Age at first symptoms (year)	1.5 [1;2]	13.3 [9;16]	15.0 [13;16]	<0.0001	<0.0001	0.04
Age at diagnosis PA (year)	3 [2.6;3.7]	16.8 [14.0;20.0]	24.0 [19.0;29.5]	<0.0001	<0.0001	<0.0001
-children and adolescents (%)	100%	71.4%	28.6%			<0.0001
Delay from symptoms to diagnosis (year)	1.6 [1;2.25]	3 [1;6]	8 [4;13]	0.06	<0.0001	0.0006
Macroadenoma (%)	76.9%	90.2%	92%	0.06	0.08	0.7
- giant adenoma (% of macroadenomas)	0.0%	10.8%	22%			0.07
Extension (%)	50.0%	77.8%	88.4%	0.02	0.001	0.07
Invasion (%)	30.0%	58.1%	66.7%	0.04	0.01	0.2
Maximal dimension of PA (mm)	18 [10;27]	25 [14;32]	24 [17;37.5]	0.3	0.2	0.65
GH levels (ng/ml)	102 [58.5;121]	23 [14.5;50]	43 [21.6;88.5]	0.01	0.09	0.05
IGF-1 levels (% UNL)	385 [285;463]	199.5 [116.8;282.7]	285 [205;403]	0.003	0.2	0.001
Prolactin co-secretion (%)	82.0%	33.3%	41%	0.02	0.01	0.4
Multimodal treatment approach (%)	46.0%	23.8%	42.7%	0.03	0.6	0.04
GH/IGF-1 control at last follow-up (%)	58.0%	61.0%	43.0%	0.7	0.02	0.03
- age when first control achieved (year)	8.0 [4;13]	17.3 [15;20]	27 [18;37]	0.006	0.0005	<0.0001
- GH/IGF-1 control ≤19 years (%)	100%	72.7%	22.7%			0.0001
- GH/IGF-1 control before final height (%)	60%	48.6%	10.9%	0.3	<0.0001	<0.0001
Long-term control (%)	41.7%	55.3%	38.4%	0.05	0.1	0.08
Follow-up period on treatment (year)	3 [0.9;8]	10.3 [4;20]	7 [3;13.5]	0.04	0.5	0.07
Hypopituitarism at follow-up (%)	75%	73%	66%	0.9	0.7	0.4

All continuous data are shown as median and IQR. PA= pituitary adenoma; MPH= mid-parental height

Discussion

In this study we report the clinical and genetic characteristics of 208 patients with pituitary gigantism due to GH hypersecretion. This, the first extensive series of patients with radiologically and hormonally proven pituitary gigantism provides insights into the disease profile of this rare disorder. Genetic or familial disease was seen in 46% of cases tested. Of tested cases, 29% had AIP mutations or deletions. Previous studies have noted that AIP mutations are associated with gigantism, either sporadic, within individual FIPA kindreds or in large historical studies (13,16–18). This high frequency of AIP mutations among pituitary gigantism patients is logical, given that AIP mutations are characteristically common among children and young adults and most frequently lead to somatotropinomas (19,20). X-LAG syndrome is a recently described form of pituitary gigantism due to chromosome Xq26.3 microduplications (14,15) and constituted 10% of the genetically studied cases in the current study. X-LAG syndrome has a particularly early age at onset, can present sporadically or as FIPA and predominantly affects females. We found no microduplication on Xq26.3 in any case diagnosed aged >5 years in our series, whereas the youngest AIP positive patient was 8 years old at diagnosis. Other genetic causes occurred less frequently; MAS, Carney complex and MEN1 comprised 7% of gigantism overall. Although pituitary gigantism can occur in both genders, it predominantly affects males (78%). This is likely due to male predominance among AIP mutated gigantism cases, as we reported previously in acromegaly (21). Among patients without AIP mutations the gender balance was heterogeneous: X-LAG syndrome cases are mainly female (14,15), whereas cases that were negative on genetic testing were predominantly male but less markedly so than the AIP mutated group. The genetically negative group comprised more than half of all cases studied. The clinical phenotype of patients with AIP mutations or X-LAG syndrome has shown to be aggressive (13,15). In this study we noted that genetically unexplained pituitary gigantism patients are even more aggressive (e.g. invasion, hormone levels, lower control rates) than AIP mutation cases. This group may be a priority for further genomic pathophysiologic studies.

We observed a high rate of pituitary tumor apoplexy in patients with gigantism (18/208 (8.7%)). Pituitary tumor apoplexy is rare (22,23), occurring in 1.6% of 1540 pituitary adenomas in one larger series (24). GH secreting adenomas account for 7.2–25.0% of such cases (25). Six cases had a genetic/inherited background: FIPA with (n=3) or without (n=1) AIP mutation; MAS (n=1) and MEN1 (n=1).

An important question regarding pituitary gigantism is whether earlier diagnosis and control of GH/IGF-1 secretion can influence final height. As this study included patients with gigantism diagnosed at any time during their growth (not only on final adult height >+2SD), we were able to address whether early recognition could limit excessive linear growth. In the group overall, the height at last follow-up was clearly in excess of MPH (11.6%; absolute difference: 20 cm) in both males and females. The median age at which linear growth ceased was 23 years, which is later than in the general population (20 years) (26). This delay permitted a longer period of growth before epiphyseal closure, a factor that was exacerbated in those with concomitant hypogonadism who had a greater final height. We found that a greater final height Z-score was determined by earlier age of onset, larger tumor size and greater GH excess. Moreover, these three features were interconnected, with younger patients developing larger tumors and higher GH secretion. Importantly, the age at which GH control was achieved had an important effect on final height. When control was achieved during the period of usual linear growth (≤19 years), the final height was lower, with a decrease in the difference between MPH and final height. These findings strongly suggest that an earlier diagnosis and a more rapid achievement of hormonal control can help to reduce final height in pituitary gigantism patients. Reducing the delay between first symptoms and the diagnosis of a pituitary adenoma relies on good awareness of the clinical features of excessive growth (including accompanying signs/symptoms), and an efficient referral process to diagnostics and treatment. This study provides scientific evidence to support improvements in disease awareness and to improve the efficiency of current diagnostic and treatment networks.

In three quarters of cases abnormal growth was the first sign/symptom reported, and it was generally established by late pre-pubertal childhood or early adolescence (median 13 years). We found that signs/symptoms were noted significantly earlier in females than in males, which led to an earlier diagnosis and shorter latency period before diagnosis. A number of factors may have contributed to this earlier diagnosis. Disease onset overlapped with the earlier pubertal growth spurt in females. The superposition of abnormal acromegaly symptoms on top of accelerated vertical growth may have led to patients seeking medical attention earlier. In addition, as tall stature even in healthy girls has long been viewed as less socially desirable (27), and this may have also contributed to an earlier recourse to medical investigation by parents and doctors. However, despite the shorter latency period in females, the difference between final height and MPH did not differ between males and females. This was probably due to the similar duration between the time of diagnosis and the time of hormonal control in the two gender groups. This highlights that earlier recognition and diagnosis needs to be accompanied by rapid therapeutic intervention to control GH/IGF-1 in order to influence final height.

Despite the young age at disease onset, pituitary adenomas were already large and most had extension and invasion at diagnosis. Elevated levels of GH and IGF-1 (with prolactin co-secretion in a third of cases) were seen at diagnosis and underpin the early and profound overgrowth seen among pituitary gigantism sufferers. Patients required multimodal treatment, with repeated surgeries and frequent use of radiotherapy. As the study was international and retrospective, not all modalities were uniformly available in all countries, particularly medical therapies like pegvisomant and SSA. Previous reports of individual cases or small series of pituitary gigantism have noted challenging disease control that required pegvisomant (28,29). More uniform early recourse to medical therapies in patients not controlled by surgery alone could theoretically improve the poor responses seen in the current cohort. However, certain genetic forms of gigantism, such as AIP mutations and X-LAG syndrome, are poorly responsive to traditional SSA, further complicating the management (15,21). Radiotherapy has a relatively slow onset of effect and may not be sufficient alone following failed surgery, as in the setting of pituitary gigantism the window to provide effective therapy and restrain overgrowth is quite narrow. Given these challenges, it would appear ideal that patients with suspected pituitary gigantism are referred to experienced centers with available multimodal therapy as soon as possible to improve chances of earlier effective disease control.

The clinical presentation included many typical disease features of adult acromegaly despite the young age of the patients. The range of signs/symptoms was mainly influenced by the duration of GH/IGF-1 hypersecretion and the delay in diagnosis. These included glucose metabolism disorders, arterial hypertension and heart disease, which are more typical of an older age group. Taking into account the poor hormonal control rate, it was not surprising that clinical symptom rates were not greatly ameliorated by surgery or on medical treatment (Supplemental table 2). Moreover hypopituitarism was diagnosed frequently in our cohort -probably due to high prevalence of macroadenomas- and rose from 25% of patients at baseline to 64% at last follow-up due to cumulative effects of treatment (i.e. surgery and radiotherapy). Given the young age of disease-associated comorbidities, relatively low control of GH/IGF-1, and the high rate of hypopituitarism, pituitary gigantism patients have significant morbidity. The impact of this morbidity on lifespan as compared with what is established in adult acromegaly is unknown (30). In our group seven patients died, all relatively young, but specific studies are required to better assess the effects of disease burden on mortality. In addition the impact of the often-dramatic physical overgrowth on quality of life in pituitary gigantism patients should be addressed.

Height is highly variable across human populations due to a variety of factors, including complex genetic influences (8,31). In addition, secular trends in anthropomorphic measures, including height, in national or regional sub-populations can lead to rapid changes over a few generations due to factors like improved nutrition (32–35). For this reason, the diagnosis of abnormal height must be made based on appropriate population norms, which ideally are country-specific and are regularly updated. We chose such normal datasets for the current study, which allowed us to classify patients with gigantism according to Z-scores for height based on their own country of origin.

This is the first study to describe the clinical, genetic and therapeutic features of pituitary gigantism in a large international cohort. Pituitary gigantism patients are predominantly males diagnosed at a young age with macroadenomas, but females have first symptoms and are diagnosed earlier than males. AIP mutations/deletions and X-LAG syndrome account for about 40% of patients tested. However, a genetic cause remains to be found in more than half of pituitary gigantism patients and these patients had aggressive disease features. Final height in gigantism was determined by an earlier age of onset, larger tumor size and greater GH excess; control of GH excess at a younger age led to a decreased final height. Treatment in patients with pituitary gigantism was complex and multimodal therapy was frequently needed. Pituitary gigantism is a challenging condition and improved management to permit rapid diagnosis and treatment would likely be aided by greater general awareness of the condition, its genetic pathophysiology and the vital role of multidisciplinary surgical and medical teams.