Phenotypic heterogeneity in pediatric pituitary stalk interruption syndrome: greater disease burden in the metabolic-onset phenotype

Abstract

Background

Pituitary stalk interruption syndrome (PSIS) is a rare congenital hypothalamic-pituitary disorder with marked clinical heterogeneity. Whether different presenting phenotypes reflect variations in disease burden and neurodevelopmental outcomes remains unclear. This study aimed to characterize distinct presenting phenotypes of children with PSIS and compare their diagnostic trajectories, endocrine and structural disease burden, and neurodevelopmental outcomes.

Methods

In this single-center retrospective study, 31 magnetic resonance imaging (MRI)-confirmed children with PSIS were classified into a systemic-onset phenotype (n=10; hypoglycemia, cholestasis, micropenis, or adrenal crisis) and a growth-predominant phenotype (n=21; short stature or growth deceleration). Clinical characteristics, pituitary hormone deficiency profiles at diagnosis and during follow-up, MRI findings, and neurodevelopmental outcomes were compared. Exploratory Spearman analyses examined potential associations between baseline clinical and endocrine characteristics and general quotient (GQ).

Results

The systemic-onset phenotype showed earlier symptom onset (0.07±0.04 vs. 2.68±0.73 years, P<0.001) but longer diagnostic lag (4.75±1.26 vs. 2.93±1.02 years, P=0.001). At diagnosis, the systemic-onset phenotype showed higher proportions of adrenocorticotropic hormone (ACTH) deficiency (80.0% vs. 33.3%, P=0.02) and luteinizing hormone/follicle-stimulating hormone (LH/FSH) deficiency (60.0% vs. 9.5%, P=0.006), a greater number of deficient pituitary axes, and lower anterior pituitary height after adjustment for age at MRI evaluation. Isolated growth hormone deficiency was present in 8 of 21 (38.1%) children with the growth-predominant phenotype but in none of those with the systemic-onset phenotype. During follow-up, at least one additional pituitary hormone deficiency was newly identified in 30.0% and 38.1% of the systemic-onset and growth-predominant groups, respectively (P>0.99), while the systemic-onset phenotype retained a greater number of deficient pituitary axes at the latest assessment (P=0.001). Neurodevelopmental outcomes were poorer, with lower GQ (P=0.007) and higher prevalence of intellectual disability (30.0% vs. 0%, P=0.03). Exploratory analyses suggested potential relationships between lower GQ and longer diagnostic lag, lower free thyroxine (FT4), lower basal cortisol, lower minimum blood glucose, and a greater number of deficient pituitary axes; these associations were exploratory and should not be interpreted causally.

Conclusions

The systemic-onset phenotype was associated with greater pituitary hormone deficiency burden at diagnosis, longer diagnostic lag, and poorer neurodevelopmental outcomes during follow-up. Initial clinical phenotype may provide a useful framework for early recognition and risk assessment, while the identification of additional pituitary hormone deficiencies during follow-up supports continued endocrine surveillance in both phenotypes.

Highlight box.

Key findings

• Children with pituitary stalk interruption syndrome (PSIS) were stratified into systemic-onset and growth-predominant phenotypes according to their predominant initial presentation.

• The systemic-onset phenotype was associated with earlier symptom onset, longer diagnostic lag, greater pituitary hormone deficiency burden at diagnosis, lower age-adjusted anterior pituitary height, and poorer neurodevelopmental outcomes. Additional hormone deficiencies were identified during follow-up in both phenotypes.

• Exploratory analyses suggested potential relationships between general quotient and diagnostic lag, minimum blood glucose, thyroid and adrenal axis function, and pituitary deficiency burden, but should not be interpreted causally.

What is known and what is new?

• PSIS is a rare congenital hypothalamic-pituitary disorder with heterogeneous presentations ranging from neonatal metabolic manifestations to childhood growth impairment. Previous studies have focused mainly on magnetic resonance imaging (MRI) findings, hormonal deficiencies, and genetics, while phenotype-related diagnostic trajectories and neurodevelopmental outcomes remain insufficiently characterized.

• This study identifies a diagnostic lag paradox in the systemic-onset phenotype, which may represent the more severely affected end of the PSIS spectrum at diagnosis. Growth-predominant presentation was not synonymous with isolated growth hormone deficiency, and additional deficiencies were identified during follow-up.

What is the implication, and what should change now?

• Persistent or recurrent neonatal hypoglycemia, particularly with cholestasis or micropenis, should prompt consideration of pituitary hormone evaluation and sellar MRI.

• Early recognition may facilitate timely hormone replacement and neurodevelopmental monitoring, while continued endocrine surveillance is warranted in both phenotypes.

• Larger multicenter prospective studies with standardized longitudinal endocrine assessments are needed to validate this framework for risk assessment and individualized management.

Introduction

Pituitary stalk interruption syndrome (PSIS) is a rare congenital hypothalamic-pituitary developmental disorder characterized by absent or attenuated pituitary stalk, ectopic posterior pituitary, and anterior pituitary hypoplasia on magnetic resonance imaging (MRI) (1). It is an important cause of congenital growth hormone (GH) deficiency and combined pituitary hormone deficiency (CPHD), and may lead to growth impairment, central hypothyroidism, adrenal insufficiency, and pubertal abnormalities (2-4). Although MRI has improved recognition of PSIS, timely diagnosis remains challenging (5).

PSIS shows marked clinical heterogeneity. Some children present in infancy with hypoglycemia, cholestasis, prolonged jaundice, micropenis, or adrenal crisis, whereas others present mainly with short stature or growth deceleration and may remain undiagnosed until late childhood or adolescence (2,6-10). However, previous studies have focused mainly on MRI findings, genetic background, and hormonal profiles (5,11), leaving unclear whether different initial presentations represent clinically meaningful phenotypes with distinct disease burden and trajectories.

Neonatal hypoglycemia, cholestasis, or adrenal insufficiency might be expected to prompt early diagnosis, but these manifestations are often attributed to sepsis, hepatitis, biliary disease, or metabolic disorders, delaying recognition of pituitary dysfunction (5,11,12). Diagnostic lag is common in PSIS, and prolonged hypoglycemia, thyroid hormone deficiency, and glucocorticoid deficiency may adversely affect brain development (13,14). Yet whether distinct initial phenotypes are associated with different diagnostic pathways, diagnostic lag patterns, and overall disease burden remains poorly characterized.

Concurrently, accumulating evidence has drawn attention to long-term neurodevelopmental outcomes in children with PSIS. Although GH deficiency has long been regarded as the most prominent feature of this condition, central hypothyroidism, central adrenal insufficiency, and recurrent hypoglycemia are considered to be associated with more profound effects on cognitive function, language development, and motor development (15-17). Existing research is largely limited to individual case reports or small case series, and knowledge regarding differences in neurodevelopmental outcomes across clinical phenotypes remains very limited (2,18).

We therefore hypothesized that PSIS comprises phenotypic subgroups with distinct disease burden and clinical trajectories. In this study, children with PSIS were classified into systemic-onset and growth-predominant phenotypes according to predominant initial presentation. We compared diagnostic processes, endocrine deficiency profiles at diagnosis and during follow-up, CPHD severity, MRI abnormalities, and neurodevelopmental outcomes to determine whether initial phenotype may support early recognition, risk stratification, and long-term management of PSIS. We present this article in accordance with the STROBE reporting checklist (available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0681/rc).

Methods

Study design and participants

This single-center retrospective study screened children evaluated at The Affiliated Children’s Hospital of Kunming Medical University between January 2021 and June 2025 for suspected PSIS or congenital hypopituitarism. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of The Affiliated Children’s Hospital of Kunming Medical University (approval No. IEC-C-008-A07-V3.0), and individual consent for this retrospective analysis was waived.

The diagnosis of PSIS was based on characteristic pituitary MRI findings, requiring at least two of the following three features:

• Absent or markedly attenuated pituitary stalk;

• Ectopic posterior pituitary;

• Anterior pituitary hypoplasia or reduced volume.

All imaging studies were independently reviewed by at least two radiologists with experience in pediatric neuroimaging.

A total of 53 potentially eligible patients were initially screened.

Inclusion criteria were:

• MRI-confirmed diagnosis of PSIS;

• Complete medical records;

• Available endocrine function assessment data;

• Available follow-up data;

• Available neurodevelopmental assessment data.

Exclusion criteria were:

• Acquired hypothalamic-pituitary injury (including tumor, radiotherapy, head trauma, or central nervous system infection);

• Confirmed chromosomal abnormality syndrome;

• Missing key clinical or endocrine data;

• Unavailable neurodevelopmental assessment data.

Of 53 screened patients, 22 were excluded: 7 without MRI-confirmed PSIS, 3 with acquired hypothalamic-pituitary injury, 2 with chromosomal or syndromic disorders, 6 with incomplete clinical or endocrine data, and 4 without neurodevelopmental assessment. Finally, 31 children were included.

Phenotypic classification

To explore the clinical heterogeneity of PSIS, participants were stratified into phenotypic subgroups according to their predominant initial presentation that prompted medical evaluation, rather than according to all clinical abnormalities present at the time of PSIS diagnosis.

Systemic-onset phenotype

Defined as an initial presentation predominantly comprising one or more of the following metabolic manifestations that prompted the initial medical evaluation:

• Neonatal or early-infantile hypoglycemia;

• Cholestasis or prolonged jaundice;

• Micropenis;

• Adrenal crisis.

Micropenis was defined as a stretched penile length (SPL) more than 2.5 standard deviations (SDs) below the mean for age, based on Chinese pediatric reference data. SPL measurements and the corresponding age at assessment were retrospectively extracted from the clinical records.

Growth-predominant phenotype

Defined as short stature or growth deceleration being the predominant initial presentation prompting medical evaluation, with at least one of the following criteria:

• Short stature, defined as a height standard deviation score (SDS) <−2.0 for age and sex; or

• Growth deceleration, defined as a height velocity SDS <−2.0 for age and sex, calculated from serial height measurements documented in the clinical records before or at the initial diagnostic evaluation.

For individuals with multiple manifestations, classification was based on the predominant clinical feature that initially prompted the family to seek medical attention and initiated the diagnostic workup. Importantly, subsequent development or recognition of growth impairment did not alter the initial phenotype classification. Thus, children classified as having the systemic-onset phenotype could also have short stature at the time of PSIS diagnosis if growth impairment developed or became apparent during the diagnostic interval.

Accordingly, 10 and 21 children were assigned to the systemic-onset and growth-predominant phenotypes groups, respectively.

Clinical data collection

Clinical data were retrospectively extracted from electronic medical records and follow-up records, including sex, birth history, age at symptom onset and diagnosis, initial manifestations and diagnoses, misdiagnosis, anthropometric data, and documented SPL, when applicable. For patients with growth-related presentations, serial height measurements before or at the initial diagnostic evaluation were reviewed to assess growth velocity. Symptom onset was defined as the age when manifestations attributable to PSIS first appeared. Confirmed diagnosis was defined as the age at which PSIS was established by endocrine evaluation together with characteristic pituitary MRI findings. As both were completed within the same diagnostic workup in this cohort, age at diagnosis did not represent separate hormonal or neuroradiological diagnostic time points. Diagnostic lag was calculated as the interval between symptom onset and confirmed diagnosis.

Endocrine data were extracted at diagnosis and at each patient’s latest available endocrine assessment during follow-up to evaluate the GH, thyroid-stimulating hormone (TSH), adrenocorticotropic hormone (ACTH), gonadotropin [luteinizing hormone (LH)/follicle-stimulating hormone (FSH)], and antidiuretic hormone (ADH) axes. A newly identified deficiency was defined as an axis classified as non-deficient at diagnosis but deficient at the latest follow-up assessment. GH deficiency was defined as a peak GH <10 ng/mL on stimulation testing, which was available for all 31 patients. Central hypothyroidism was defined as free thyroxine (FT4) below the age-specific reference range with an inappropriately low, normal, or mildly elevated TSH. ACTH deficiency was determined from clinical evidence of central adrenal insufficiency together with morning basal cortisol and corresponding ACTH concentrations; ACTH stimulation testing was not performed.

Gonadotropin deficiency was assessed according to age and pubertal stage using pubertal development, Tanner stage, basal LH/FSH, and sex steroid levels. In children of appropriate pubertal age, low or inappropriately normal gonadotropins with low sex steroids and absent or impaired pubertal development were considered indicative of deficiency; low basal gonadotropins alone in prepubertal children were not diagnostic. Gonadotropin-releasing hormone (GnRH) stimulation testing was not performed. ADH deficiency was determined from documented clinical and biochemical evidence of central diabetes insipidus, including persistent polyuria/polydipsia and available serum and urinary biochemical findings; dynamic ADH testing was not performed. Cholestasis was defined as direct (conjugated) bilirubin >1.0 mg/dL (>17 μmol/L).

Pituitary MRI was evaluated for absent or attenuated pituitary stalk, ectopic posterior pituitary, and anterior pituitary hypoplasia. Anterior pituitary height was measured as the maximum craniocaudal height on the mid-sagittal plane. Measurements were reviewed by two radiologists, with discrepancies resolved by consensus.

Neurodevelopment was assessed once during follow-up after PSIS diagnosis and initiation of clinically indicated hormone replacement therapy using the Griffiths Development Scales-Chinese (GDS-C) (19). The interval from treatment initiation to assessment was not prespecified because of the retrospective design. General quotient (GQ) was the primary neurodevelopmental outcome, and intellectual disability, motor developmental delay, and language developmental delay were recorded according to standardized clinical criteria.

Outcome measures

The primary outcomes were: phenotype-associated differences in endocrine and structural disease burden between phenotypic groups, including:

• Hormonal deficiency profiles at diagnosis;

• Number of deficient pituitary axes at diagnosis;

• MRI structural abnormality characteristics.

Secondary outcomes included:

• Differences in diagnostic pathways;

• Diagnostic lag;

• Changes in pituitary hormone deficiency status from diagnosis to the latest available endocrine assessment;

• Neurodevelopmental outcomes;

• Exploratory analysis of factors associated with GQ.

Statistical analysis

Statistical analyses were performed using IBM SPSS Statistics (version 27.0; IBM Corp., Armonk, NY, USA), GraphPad Prism (version 10.0; GraphPad Software, San Diego, CA, USA), and Python (version 3.11).

Continuous variables were summarized as mean ± SD or median [interquartile range (IQR)], as appropriate, and categorical variables as frequencies and percentages. Comparisons between the two independent phenotype groups were performed using the Mann-Whitney U test for continuous variables and Fisher’s exact test for categorical variables. For longitudinal endocrine analyses, newly identified pituitary hormone deficiencies during follow-up were summarized as frequencies and percentages. Between-phenotype differences were assessed using Fisher’s exact test. Change in the number of deficient pituitary axes was calculated as the number at the latest available assessment minus that at diagnosis and compared between phenotypes using the Mann-Whitney U test.

Analyses of micropenis were restricted to boys. Spearman correlation assessed associations between GQ and diagnostic lag, minimum blood glucose, FT4 and basal cortisol measured at the diagnostic evaluation, the number of deficient pituitary axes at diagnosis, and age at neurodevelopmental assessment. Bootstrap resampling with 5,000 iterations estimated 95% confidence intervals (CIs) for Spearman coefficients. All tests were two-sided, with P<0.05 considered statistically significant. Given the small sample size, correlation analyses were exploratory, were not adjusted for multiple comparisons, and should not be interpreted causally.

Results

Distinct clinical trajectories define two PSIS phenotypes

A total of 31 children with MRI-confirmed PSIS and complete clinical records were included. According to predominant initial presentation, 10 were assigned to the systemic-onset phenotype group and 21 to the growth-predominant phenotype group. The study flowchart and classification strategy are shown in Figure 1.

Figure 1.

Study flowchart and phenotypic classification of patients with PSIS. Flowchart illustrating patient screening, eligibility assessment, exclusion process, and phenotypic classification. A total of 53 children identified from the pediatric endocrinology database were screened, of whom 31 met the eligibility criteria and were included in the final analytic cohort. Patients were categorized according to their predominant initial presentation into a systemic-onset phenotype (hypoglycemia, cholestasis, micropenis, or adrenal crisis) or a growth-predominant phenotype (short stature or growth deceleration). Comparative analyses were subsequently performed to evaluate diagnostic pathways, endocrine burden, MRI characteristics, and neurodevelopmental outcomes. MRI, magnetic resonance imaging; PSIS, pituitary stalk interruption syndrome.

The two groups did not differ significantly with respect to sex distribution, gestational age, birth weight, rate of breech delivery, or rate of perinatal asphyxia (Table 1), suggesting that the two phenotypes were not attributable to differences in perinatal background. No cryptorchidism was documented in any child in the cohort.

Table 1. Baseline demographic and diagnostic characteristics according to presenting phenotype.

Variable	Systemic-onset phenotype	Growth-predominant phenotype	P
Sex			0.70
Female	3 (30.0)	9 (42.9)
Male	7 (70.0)	12 (57.1)
Gestational age, weeks	39.00 [38.00, 40.00]	39.00 [38.00, 40.00]	0.91
Birth weight, kg	3.11±0.26	3.07±0.30	0.69
Breech delivery			0.42
Yes	4 (40.0)	5 (23.8)
No	6 (60.0)	16 (76.2)
Perinatal asphyxia			>0.99
Yes	1 (10.0)	4 (19.0)
No	9 (90.0)	17 (81.0)
Age at symptom onset, years	0.07±0.04	2.68±0.73	<0.001
Age at diagnosis, years	4.81±1.24	5.61±0.85	0.12
Diagnostic lag, years	4.75±1.26	2.93±1.02	0.001
Initial presentation			<0.001
Hypoglycemia	2 (20.0)	–
Hypoglycemia + cholestasis + seizure	3 (30.0)	–
Hypoglycemia + micropenis + seizure	3 (30.0)	–
Hypoglycemia + micropenis	1 (10.0)	–
Hypoglycemia + cholestasis + micropenis	1 (10.0)	–
Short stature	–	15 (71.4)
Growth deceleration	–	5 (23.8)
Growth deceleration + occasional hypoglycemia	–	1 (4.8)
Initial misdiagnosis	–	1 (4.8)	<0.001
No misdiagnosis	–	15 (71.4)
Neonatal hepatitis	3 (30.0)	–
Epilepsy	2 (20.0)	–
Sepsis	2 (20.0)	–
Cholestatic liver disease	2 (20.0)
Hypoglycemia	1 (10.0)
Malnutrition	–	5 (23.8)
Constitutional growth delay	–	1 (4.8)
Height, SDS	−2.61±0.94	−4.07±0.86	<0.001
BMI, SDS	−0.62±0.47	−0.52±0.35	0.58

Data are presented as mean ± SD, median [IQR], or n (%), as appropriate. Gestational age is expressed in weeks, birth weight in kilograms, and age-related variables in years. Phenotypic classification was based on the predominant initial clinical presentation that prompted medical evaluation, rather than on all clinical abnormalities present at the time of PSIS diagnosis. Height SDS was assessed at the time of diagnosis; therefore, patients classified as having the systemic-onset phenotype could also exhibit short stature by the time of diagnostic evaluation. BMI, body mass index; IQR, interquartile range; PSIS, pituitary stalk interruption syndrome; SD, standard deviation; SDS, standard deviation score.

In addition, a history of seizures was more prevalent in the systemic-onset phenotype (60.0% vs. 19.0%, P=0.04). Micropenis was also more frequently observed in the systemic-onset phenotype than in the growth-predominant phenotype (50.0% vs. 9.5%, P=0.045).

Despite earlier clinical presentation, children with the systemic-onset phenotype were not diagnosed earlier than those with the growth-predominant phenotype. The two groups showed distinct diagnostic trajectories. At the individual level, patients with the systemic-onset phenotype mainly presented in infancy with hypoglycemia, cholestasis, micropenis, and related systemic manifestations, whereas those with the growth-predominant phenotype presented primarily with short stature or growth deceleration (Figure 2A). Age at symptom onset was significantly earlier in the systemic-onset phenotype than in the growth-predominant phenotype (0.07±0.04 vs. 2.68±0.73 years, P<0.001), whereas age at confirmed diagnosis was similar between groups (4.81±1.24 vs. 5.61±0.85 years, P=0.12), resulting in a longer diagnostic lag in the systemic-onset phenotype (4.75±1.26 vs. 2.93±1.02 years, P=0.001) (Table 1).

Figure 2.

Distinct diagnostic trajectories of the systemic-onset and growth-predominant phenotypes. (A) Individual-level timelines illustrating age at symptom onset, age at diagnosis, and diagnostic lag in each patient. Each horizontal bar represents one patient. Colors indicate phenotypic classification. (B) Comparison of age at symptom onset, age at diagnosis, and diagnostic lag between phenotypes. Group comparisons were performed using the Mann-Whitney U test. NS, P≥0.05; ***, P<0.001. NS, not significant.

This disparity is evident in the individual diagnostic timelines in Figure 2. Most children with the systemic-onset phenotype had a gap of several years between symptom onset and diagnosis, whereas those with the growth-predominant phenotype, despite later presentation, more often underwent endocrine evaluation soon after growth concerns emerged. Consistently, the systemic-onset phenotype showed earlier symptom onset, longer diagnostic lag, and similar age at diagnosis compared with the growth-predominant phenotype (Figure 2B).

Initial diagnostic labels provided additional context for this diagnostic lag. Children with the systemic-onset phenotype were frequently assigned non-endocrine diagnoses before PSIS recognition, most commonly neonatal hepatitis, cholestatic liver disease, epilepsy, and sepsis. In contrast, 71.4% of patients in the growth-predominant phenotype group had no clearly erroneous initial diagnosis (Table 1). These findings indicate different recognition pathways in routine practice: growth-related manifestations more readily prompted endocrine evaluation, whereas metabolic presentations were often interpreted within hepatologic, neurologic, or infectious frameworks and were associated with later recognition of hypothalamic-pituitary disease. Together, these observations illustrate what we term a “diagnostic lag paradox” in PSIS: despite earlier presentation, children with the systemic-onset phenotype experienced a more prolonged pathway to diagnosis.

The systemic-onset phenotype is associated with greater endocrine and structural disease burden

We subsequently compared pituitary function and MRI structural characteristics between the two groups.

Analysis of pituitary hormonal deficiencies demonstrated a greater endocrine deficiency burden in patients with the systemic-onset phenotype (Table 2, Figure 3A). Importantly, the growth-predominant phenotype was not synonymous with isolated GHD: although 8 of 21 (38.1%) children in this group had isolated GHD at diagnosis, the remaining patients had additional pituitary hormone deficiencies. In contrast, none of the children with the systemic-onset phenotype had isolated GHD at diagnosis (0% vs. 38.1%, P=0.03). Compared with the growth-predominant phenotype, the systemic-onset phenotype showed higher proportions of ACTH deficiency (80.0% vs. 33.3%, P=0.02) and LH/FSH deficiency (60.0% vs. 9.5%, P=0.006). TSH deficiency was numerically more frequent in the systemic-onset phenotype (80.0% vs. 47.6%), although the difference was not statistically significant (P=0.13). GH deficiency was present in all patients in both groups (100.0% vs. 100.0%).

Table 2. Endocrine deficiency profile and burden at diagnosis according to presenting phenotype.

Variable	Systemic-onset phenotype (n=10)	Growth-predominant phenotype (n=21)	P
GH deficiency	10 (100.0)	21 (100.0)	–
TSH deficiency	8 (80.0)	10 (47.6)	0.13
ACTH deficiency	8 (80.0)	7 (33.3)	0.02
LH/FSH deficiency	6 (60.0)	2 (9.5)	0.006
ADH deficiency	0 (0)	0 (0)	–
Number of deficient pituitary axes	3 [3–4]	2 [1–2]	0.001
Isolated GHD	0 (0)	8 (38.1)	0.03

Data are presented as n (%) or median [IQR], as appropriate. Categorical variables were compared using Fisher’s exact test, and the number of deficient pituitary axes was compared using the Mann-Whitney U test. Isolated GHD was defined as GH deficiency without any additional pituitary hormone deficiency at diagnosis; patients not meeting this definition had multiple pituitary hormone deficiencies. No P value was calculated for GH or ADH deficiency because there was no between-patient variation in these variables. ACTH, adrenocorticotropic hormone; ADH, antidiuretic hormone; FSH, follicle-stimulating hormone; GH, growth hormone; GHD, growth hormone deficiency; IQR, interquartile range; LH, luteinizing hormone; TSH, thyroid-stimulating hormone.

Figure 3.

Endocrine deficiency burden and pituitary structural characteristics according to presenting phenotype. (A) Prevalence of pituitary hormone deficiencies in the systemic-onset and growth-predominant phenotypes. (B) Comparison of the number of deficient pituitary axes between phenotypes. (C) Frequency of characteristic MRI abnormalities, including absent pituitary stalk, attenuated pituitary stalk, ectopic posterior pituitary, and anterior pituitary hypoplasia. (D) Comparison of age-adjusted anterior pituitary height between phenotypes. Data are presented as median (IQR) or percentages, as appropriate. Continuous variables were compared using the Mann-Whitney U test, and categorical variables using Fisher’s exact test, as appropriate. For anterior pituitary height, between-group differences were assessed after adjustment for age at MRI evaluation. NS, P≥0.05; *, P<0.05; **, P<0.01. ACTH, adrenocorticotropic hormone; ADH, antidiuretic hormone; AP, anterior pituitary; FSH, follicle-stimulating hormone; GH, growth hormone; IQR, interquartile range; LH, luteinizing hormone; MRI, magnetic resonance imaging; NS, not significant; PP, posterior pituitary; TSH, thyroid-stimulating hormone .

Consistent with the hormonal deficiency profile, patients with the systemic-onset phenotype had a significantly greater number of deficient pituitary axes than those with the growth-predominant phenotype {median [IQR] 4 [3–4] vs. 2 [1–3], P<0.001; Figure 3B}.

MRI structural abnormalities were common in both phenotypes. The frequencies of absent pituitary stalk, attenuated pituitary stalk, ectopic posterior pituitary, and anterior pituitary hypoplasia did not differ significantly between groups (all P>0.05; Figure 3C). After adjustment for age at MRI evaluation, anterior pituitary height remained significantly lower in the systemic-onset phenotype than in the growth-predominant phenotype (P<0.01; Figure 3D).

Collectively, the systemic-onset phenotype was characterized by a higher prevalence of ACTH and LH/FSH deficiencies, a greater number of deficient pituitary axes, and lower anterior pituitary height after adjustment for age at MRI evaluation compared with the growth-predominant phenotype.

Changes in pituitary hormone deficiencies during follow-up

Pituitary hormone deficiencies were further evaluated at the latest available endocrine assessment to assess changes in deficiency status during follow-up (Table 3). At the latest available endocrine assessment, the systemic-onset phenotype continued to show a greater overall endocrine deficiency burden than the growth-predominant phenotype. LH/FSH deficiency remained more frequent in the systemic-onset phenotype (60.0% vs. 9.5%, P=0.006), and the number of deficient pituitary axes was greater in the systemic-onset phenotype {median [IQR], 3.5 [3–4] vs. 2 [2–3], P=0.001}. Differences in TSH deficiency (100.0% vs. 66.7%, P=0.07) and ACTH deficiency (90.0% vs. 52.4%, P=0.055) did not reach statistical significance.

Table 3. Changes in pituitary hormone deficiencies during follow-up according to presenting phenotype.

Variable	Systemic-onset phenotype (n=10)	Growth-predominant phenotype (n=21)	P
Pituitary hormone deficiencies at latest follow-up
GH deficiency	10 (100.0)	21 (100.0)	–
TSH deficiency	10 (100.0)	14 (66.7)	0.07
ACTH deficiency	9 (90.0)	11 (52.4)	0.055
LH/FSH deficiency	6 (60.0)	2 (9.5)	0.006
ADH deficiency	0 (0)	0 (0)	–
Number of deficient pituitary axes	3.5 [3–4]	2 [2–3]	0.001
Changes during follow-up
≥1 newly identified pituitary hormone deficiency	3 (30.0)	8 (38.1)	>0.99
Newly identified TSH deficiency	2 (20.0)	4 (19.0)	>0.99
Newly identified ACTH deficiency	1 (10.0)	4 (19.0)	>0.99
Newly identified LH/FSH deficiency	0 (0)	0 (0)	–
Newly identified ADH deficiency	0 (0)	0 (0)	–
Change in number of deficient pituitary axes	0 [0–0.75]	0 [0–1]	0.68

Data are presented as n (%) or median [IQR], as appropriate. Endocrine status at follow-up was based on the latest available endocrine assessment for each patient. A newly identified pituitary hormone deficiency was defined as an axis not classified as deficient at diagnosis but subsequently documented as deficient during follow-up. Between-group comparisons were performed using Fisher’s exact test for categorical variables and the Mann-Whitney U test for the number and change in number of deficient pituitary axes. No P value was calculated when there was no between-patient variation. Baseline endocrine deficiency status at diagnosis is presented in Table 2. ACTH, adrenocorticotropic hormone; ADH, antidiuretic hormone; FSH, follicle-stimulating hormone; GH, growth hormone; IQR, interquartile range; LH, luteinizing hormone; TSH, thyroid-stimulating hormone.

During follow-up, at least one additional pituitary hormone deficiency was newly identified in 3 of 10 (30.0%) children with the systemic-onset phenotype and 8 of 21 (38.1%) children with the growth-predominant phenotype (P>0.99). In the growth-predominant phenotype, newly identified deficiencies included the TSH and ACTH axes, indicating that deficiencies involving additional pituitary axes were identified during follow-up in some patients with a growth-predominant initial presentation. The change in the number of deficient pituitary axes from diagnosis to the latest available assessment did not differ significantly between the two phenotypes {median [IQR], 0 [0–0.75] vs. 0 [0–1], P=0.68}.

These longitudinal observations indicate that additional pituitary hormone deficiencies were identified during follow-up in both phenotypes, including among children whose initial presentation was growth-predominant. Although the systemic-onset phenotype retained a greater endocrine deficiency burden at the latest assessment, the extent of newly identified endocrine deficiencies during follow-up did not differ significantly between the two groups.

Neurodevelopmental outcomes differ substantially between phenotypes

Neurodevelopmental assessments, performed during follow-up after initiation of clinically indicated hormone replacement therapy, revealed significant differences between the two phenotypes (Table 4).

Table 4. Neurodevelopmental outcomes according to presenting phenotype.

Variable	Systemic-onset phenotype	Growth-predominant phenotype	P
Age at neurodevelopmental assessment, years	5.45 [4.40–6.33]	5.90 [5.00–6.60]	0.22
GQ	75.00 [68.75–84.25]	92.00 [90.00–95.00]	0.007
Intellectual disability			0.03
Yes	3 (30.0)	0
No	7 (70.0)	21 (100.0)
Motor developmental delay			0.03
Yes	4 (40.0)	1 (4.8)
No	6 (60.0)	20 (95.2)
Language developmental delay			0.02
Yes	6 (60.0)	3 (14.3)	0.02
No	4 (40.0)	18 (85.7)
Seizure			0.04
Yes	6 (60.0)	4 (19.0)
No	4 (40.0)	17 (81.0)
Micropenis†			0.045
Yes	5 (50.0)	2 (9.5)
No	2 (20.0)	10 (47.6)
Not applicable	3 (30.0)	9 (42.9)
Cryptorchidism†
Yes	0	0	>0.99†
No	7 (70.0)	12 (57.1)
Not applicable	3 (30.0)	9 (42.9)

Continuous variables are presented as median [IQR] and categorical variables as n (%). Continuous variables were compared using the Mann-Whitney U test. Categorical variables were compared using Fisher’s exact test. ^†, analyses of micropenis and cryptorchidism were restricted to male participants; percentages are shown using the entire phenotype group as the denominator for descriptive purposes, and P values were calculated among male participants only. GQ, general quotient; IQR, interquartile range.

Ages at neurodevelopmental assessment were comparable between groups {median [IQR], 5.45 [4.40–6.33] vs. 5.90 [5.00–6.60] years, P=0.22}. However, children with the systemic-onset phenotype exhibited significantly lower GQ scores than those with the growth-predominant phenotype {median [IQR], 75.00 [68.75–84.25] vs. 92.00 [90.00–95.00], P=0.007}. Further analysis demonstrated that intellectual disability was observed in 30.0% of patients with the systemic-onset phenotype and in none of those with the growth-predominant phenotype (P=0.03). Motor developmental delay (40.0% vs. 4.8%, P=0.03) and language developmental delay (60.0% vs. 14.3%, P=0.02) were likewise more common in the systemic-onset phenotype.

Collectively, the systemic-onset phenotype was characterized by lower GQ scores and a higher prevalence of intellectual disability, motor and language developmental delay, and seizures compared with the growth-predominant phenotype.

Factors associated with cognitive and developmental outcomes

To further explore factors potentially associated with neurodevelopmental outcomes, we conducted an exploratory correlation analysis in the overall cohort using clinical and endocrine variables obtained at diagnosis, together with age at neurodevelopmental assessment (Figure 4).

Figure 4.

Exploratory correlations of diagnostic delay, endocrine measures, and pituitary deficiency burden with neurodevelopmental outcome in children with PSIS. Scatterplots showing exploratory correlations between GQ and (A) diagnostic lag, (B) FT4 concentration, (C) basal cortisol concentration, (D) minimum recorded blood glucose level, (E) number of deficient pituitary axes, and (F) age at neurodevelopmental assessment. Spearman rank correlation coefficients (ρ) and corresponding P values are shown in each panel. For the analyses of FT4 and basal cortisol, the entire cohort was included irrespective of endocrine deficiency status; therefore, the observed correlations may partly reflect differences between patients with normal and abnormal endocrine function rather than continuous associations within patients with the corresponding endocrine deficiency. These analyses were exploratory and were not intended to establish independent, within-deficiency, or causal relationships. Shaded areas represent 95% confidence intervals for the fitted regression lines. CI, confidence interval; FT4, free thyroxine; GQ, general quotient; PSIS, pituitary stalk interruption syndrome.

In the exploratory analysis, GQ was negatively correlated with diagnostic lag (ρ=−0.686, P<0.001) (Figure 4A). FT4 concentration (ρ=0.535, P=0.002) and basal cortisol level (ρ=0.630, P<0.001) showed positive correlations with GQ (Figure 4B,4C). Because these analyses included patients with both normal and abnormal endocrine function, the observed correlations may partly reflect differences between these endocrine status groups rather than continuous associations within patients with the corresponding endocrine deficiency.

Minimum recorded blood glucose level showed a positive correlation with GQ (ρ=0.552, P=0.001) (Figure 4D). The number of deficient pituitary axes showed a negative correlation with GQ (ρ=−0.701, P<0.001) (Figure 4E). In contrast, no correlation was evident between age at neurodevelopmental assessment and GQ (ρ=−0.071, P=0.70) (Figure 4F).

Overall, these exploratory analyses suggested potential relationships of GQ with diagnostic lag, FT4, basal cortisol, minimum blood glucose level, and the number of deficient pituitary axes. In particular, the associations involving FT4 and basal cortisol should not be interpreted as evidence of continuous relationships within patients with the corresponding endocrine deficiencies. Given the small sample size and the absence of correction for multiple comparisons, these findings should be interpreted cautiously and require confirmation in larger cohorts.

Discussion

PSIS is a rare congenital hypothalamic-pituitary disorder with heterogeneous manifestations, ranging from neonatal crises to childhood growth impairment and later endocrine dysfunction. Because of its rarity and nonspecific early features, the relationship between initial phenotype and long-term outcomes remains poorly defined. Previous studies have mainly addressed MRI findings, hormonal deficits, and genetic mechanisms, with limited attention to diagnostic trajectories, disease burden, and neurodevelopmental outcomes across phenotypes (5,20). In this cohort of 31 MRI-confirmed children with PSIS, we found that the systemic-onset phenotype was associated with greater endocrine burden at diagnosis, lower anterior pituitary height after adjustment for age at MRI evaluation, longer diagnostic lag, and poorer neurodevelopmental outcomes. Additional pituitary hormone deficiencies were identified during follow-up in both phenotypes, while the systemic-onset phenotype retained a greater overall endocrine deficiency burden at the latest available assessment. These findings suggest that initial presentation may provide a practical framework for phenotypic stratification and risk assessment.

Particularly noteworthy was the observation that children with the systemic-onset phenotype presented substantially earlier than those with the growth-predominant phenotype (0.07±0.04 vs. 2.68±0.73 years), yet reached a similar age at diagnosis and experienced a significantly longer diagnostic lag (4.75±1.26 vs. 2.93±1.02 years). This “diagnostic lag paradox” is one of the most distinctive findings of this study. Conventional reasoning would suggest that earlier symptom onset should translate into earlier engagement with the healthcare system and thus earlier diagnosis. In PSIS, however, this assumption may not hold because the earliest manifestations are often nonspecific and may not immediately prompt endocrine evaluation. Children with the systemic-onset phenotype predominantly present with neonatal hypoglycemia, cholestasis, prolonged jaundice, and micropenis. As observed in our cohort, these manifestations were frequently associated with initial diagnoses such as neonatal hepatitis, cholestatic liver disease, epilepsy, sepsis, or other non-endocrine conditions, thereby obscuring the underlying hypothalamic-pituitary disorder and delaying definitive diagnosis (5,12). This pattern may reflect limited early recognition of congenital hypopituitarism and delays in referral for endocrine evaluation; however, the retrospective nature of this study precluded determination of the specific reasons why endocrine investigations were not performed at the initial presentation in individual patients.

Clinically, diagnostic delay may reflect a mismatch between symptom severity and recognition pathway. Growth deceleration is persistent and quantifiable, making endocrine referral more likely. In contrast, neonatal metabolic crises may be transient, and once acute symptoms resolve, underlying endocrine deficiencies may remain unrecognized (3,18,21). Therefore, persistent or recurrent neonatal hypoglycemia, particularly when accompanied by micropenis or cholestasis, should prompt consideration of congenital hypopituitarism and early pituitary hormone evaluation, with sellar MRI when clinically indicated, rather than evaluation limited to infectious, hepatobiliary, or metabolic causes.

Beyond diagnostic delay, endocrine profiles suggested greater underlying disease severity in the systemic-onset phenotype. Although GH deficiency was universal, ACTH and LH/FSH deficiencies were more frequent in the systemic-onset phenotype, which also had a greater number of deficient pituitary axes at diagnosis. Nevertheless, the growth-predominant phenotype should not be interpreted as representing isolated GHD, because only a subset of these children had isolated GHD at diagnosis, whereas others already had deficiencies involving additional pituitary axes. Notably, all children with the systemic-onset phenotype had multiple pituitary hormone deficiencies at diagnosis, whereas a subset of those with the growth-predominant phenotype presented with isolated GHD, further supporting a greater initial endocrine burden in the systemic-onset phenotype. The more severe height deficits in the growth-predominant phenotype likely reflect, at least in part, a classification and referral effect, because short stature or growth deceleration defined this phenotype and triggered medical evaluation. Moreover, height SDS represents the cumulative outcome of linear growth over time and is influenced by multiple factors, whereas GH peak and IGF-1 SDS obtained during diagnostic evaluation represent biochemical measures at a specific time point. Therefore, the apparently lower height SDS in the growth-predominant phenotype despite relatively higher GH peak and IGF-1 SDS should not be interpreted as contradictory or as indicating greater overall endocrine disease severity. Differences in age, duration and timing of growth impairment, and nutritional status may also have contributed to the observed auxological differences; however, their relative contributions could not be determined in this small retrospective cohort. Overall, considering the broader pituitary hormone deficiency profile rather than height SDS alone, the systemic-onset phenotype may represent the more severely affected end of the PSIS spectrum rather than a completely distinct disease subtype.

Longitudinal endocrine observations showed that additional pituitary hormone deficiencies were identified after diagnosis in both phenotypes, including TSH and ACTH deficiencies among children with the growth-predominant phenotype. These findings indicate that a growth-predominant presentation does not imply persistent isolated involvement of the GH axis. At the latest available endocrine assessment, the systemic-onset phenotype continued to show a greater number of deficient pituitary axes. However, neither the proportion of patients with newly identified deficiencies nor the change in the number of deficient axes differed significantly between phenotypes. Thus, the longitudinal findings do not indicate more rapid endocrine progression in the systemic-onset phenotype; rather, they suggest that its greater endocrine burden was already evident at diagnosis and remained apparent at the latest available endocrine assessment. Importantly, the identification of additional deficiencies in some children with the growth-predominant phenotype underscores the need for continued surveillance of multiple pituitary axes even when the initial presentation is dominated by growth impairment.

The neurodevelopmental findings should be interpreted in the context of both the underlying endocrine phenotype and the timing of diagnosis and treatment. Deficiencies of the thyroid and adrenal axes may adversely affect brain development, particularly during the critical period from late fetal life to the first two postnatal years, when myelination, neuronal migration, and synaptic network formation depend on adequate hormonal regulation (22,23). Thus, combined impairment of these endocrine axes may influence not only metabolic stability but also neurodevelopmental trajectory. Importantly, among children with central hypothyroidism, the longer diagnostic lag observed in the systemic-onset phenotype may also have prolonged the period of unrecognized and untreated thyroid hormone deficiency, potentially contributing to the poorer neurodevelopmental outcomes observed in this group.

Exploratory correlation analyses suggested potential relationships between GQ and diagnostic lag, FT4, basal cortisol, minimum blood glucose, and the number of deficient pituitary axes; however, these associations may also be influenced by differences in the timing of diagnosis and initiation of hormone replacement therapy. These findings should therefore be interpreted as hypothesis-generating rather than as evidence of independent or causal relationships between endocrine status at diagnosis and neurodevelopmental status assessed during follow-up. Given the small sample size, retrospective design, and absence of adjustment for multiple comparisons, these exploratory findings require confirmation in larger studies.

Several limitations should be acknowledged. First, the small sample size and single-center retrospective design limited statistical power, particularly for subgroup comparisons and exploratory correlation analyses, and precluded robust multivariable adjustment. Second, endocrine characterization was based on data obtained during routine clinical care; although the major pituitary axes were systematically classified using predefined clinical and biochemical criteria, dynamic testing of the ACTH, gonadotropin, and ADH axes was not available in this cohort. Third, endocrine reassessments during follow-up were not performed at prespecified uniform intervals, limiting interpretation of the timing and rate of newly identified deficiencies. In addition, differences in follow-up duration may have resulted in unequal opportunities for newly emerging or newly recognized deficiencies to be detected across patients. Fourth, neurodevelopment was assessed only once during follow-up after treatment initiation; therefore, changes in neurodevelopment from diagnosis to follow-up and the effects of hormone replacement therapy on GQ could not be directly evaluated. Finally, nutritional status and other determinants of linear growth were not systematically evaluated. These limitations warrant cautious interpretation of the findings and validation in larger, multicenter prospective cohorts with standardized longitudinal endocrine assessments.

Conclusions

In this single-center retrospective study of 31 children with PSIS, two presenting phenotypes were characterized according to predominant initial presentation. Compared with the growth-predominant phenotype, the systemic-onset phenotype was associated with earlier symptom onset, longer diagnostic lag, greater pituitary hormone deficiency burden at diagnosis, lower age-adjusted anterior pituitary height, and poorer neurodevelopmental outcomes. Additional pituitary hormone deficiencies were identified during follow-up in both phenotypes, while the systemic-onset phenotype retained a greater endocrine burden at the latest available assessment. Exploratory analyses suggested potential relationships between GQ and diagnostic lag, minimum blood glucose, thyroid and adrenal axis function, and pituitary deficiency burden, although these findings should not be interpreted causally.

Overall, the systemic-onset phenotype may represent the more severe end of the PSIS spectrum at diagnosis. Early recognition may facilitate timely endocrine evaluation, hormone replacement, and neurodevelopmental monitoring. Continued endocrine surveillance is warranted in both phenotypes, and larger prospective studies with standardized longitudinal assessments are needed to validate this phenotypic framework.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of The Affiliated Children’s Hospital of Kunming Medical University (Approval No. IEC-C-008-A07-V3.0), and individual consent for this retrospective analysis was waived.

Data Sharing Statement

Available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0681/dss

tp-15-9-363-dss.pdf ^{(93.1KB, pdf)}

DOI: 10.21037/tp-2026-0681