Increased pituitary volume in subjects at risk for psychosis and patients with first-episode schizophrenia




Enlarged pituitary gland has been reported in schizophrenia, possibly reflecting hypothalamic–pituitary–adrenal hyperactivity. The aim of the present study was to examine whether individuals at risk of psychosis also have similar changes.


Magnetic resonance imaging was used to examine the pituitary volume in 22 individuals with at-risk mental state (ARMS; 11 male, 11 female), 64 first-episode patients with schizophrenia (FESz; 37 male, 27 female), and 86 healthy controls. The control subjects were divided into age- and gender-matched controls for ARMS (11 male, 11 female) and FESz (37 male, 27 female).


Both the ARMS and FESz groups had a larger pituitary volume compared with matched controls, but no difference was found between the ARMS and FESz subjects. There was no association between the pituitary volume and clinical variables (symptommeasures at scanning, daily dosage or duration of antipsychotic medication) in either clinical group. The pituitary volume did not differ significantly between the ARMS individuals who later developed schizophrenia (n = 5) and those who did not (n = 17). The pituitary volume was larger in women than in men for all diagnostic groups.


The finding of increased pituitary volume in both ARMS and FESz subjects may reflect a common vulnerability to stress in early psychosis. Further work in a larger ARMS sample is required to examine the possible relationship between pituitary volume and emergence of psychosis.

HYPOTHALAMIC–PITUITARY–ADRENAL (HPA) axis hyperactivity is thought to reflect stress-related hormonal dysregulation and has been described in schizophrenia.[1, 2] Although not consistently replicated,[3] previous magnetic resonance imaging (MRI) studies have generally demonstrated enlarged pituitary volume[4-6] with ongoing expansion[7, 8] early in the course of schizophrenia, presumably reflecting activation of the hormonal stress response. The patients may also exhibit pituitary atrophy during later courses,[9-11] possibly as a result of prolonged HPA activation.[12] Interestingly, recent neuroendocrine findings in clinical subjects at high risk for developing psychosis (i.e. at-risk mental state; ARMS[13]), such as the association of cortisol level with prodromal or psychotic symptoms[14-17] as well as with progression to psychosis,[18] suggest that HPA axis dysfunction may pre-date the onset of psychosis in at least some individuals.[19]

In contrast to these hormonal investigations, there have been only a few MRI studies addressing pituitary volume changes prior to psychosis onset and the results have been inconsistent. In the first MRI study of the pituitary gland in clinical high-risk subjects, Garner et al. found no significant volume difference between the ARMS subjects (as a whole or those who later developed psychosis) and controls, but pituitary enlargement was associated with later transition to psychosis (predominantly affective psychosis).[4] They also examined the possible relationship of the pituitary volume to anxiety/depressive or psychotic symptoms, but found no significant results. Thompson et al. showed that pituitary volume in ARMS did not correlate with the experience of stressful events, plasma cortisol level, or clinical symptoms, but that study lacked a healthy comparison group.[17] A recent study by Büschlen et al. did not replicate a significant difference in the pituitary volume between ARMS with and without transition,[4] but their data (controls < ARMS without later transition < ARMS with transition and first-episode psychosis)[20] were in line with hypothesized pituitary enlargement with the emergence of psychosis. Thus, it remains unclear from the current evidence whether these high-risk subjects have significant pituitary volume changes as compared with controls and whether their pituitary volume is related to clinical characteristics.

The present MRI study investigated the pituitary volume in subjects with ARMS and first-episode schizophrenia (FESz) compared with age- and gender-matched healthy controls. On the basis of previous MRI and neuroendocrine findings suggesting HPA hyperactivity prior to the onset of overt psychosis,[19] we predicted that both ARMS and FESz subjects would have increased pituitary volume compared with matched controls. We also explored the relationship between the pituitary volume and clinical characteristics (e.g. symptom severity, later transition into psychosis, and antipsychotic medication) in these participants.



Twenty-two ARMS subjects were recruited from the Consultation Support Service in Toyama (CAST), which was launched in 2006 as a specialized clinical setting to study and treat young people (aged 15–30 years) at risk for developing psychosis.[21] The ARMS subjects, who had no previous episode of overt psychosis and no clear diagnosis of major depression or borderline personality disorder, were diagnosed according to the Comprehensive Assessment of At Risk Mental States (CAARMS);[13] inclusion into the study required one or more of (i) attenuated psychotic symptoms defined by subthreshold intensity or frequency (n = 21); (ii) brief limited intermittent psychotic symptoms with spontaneous resolution (n = 2); and/or (iii) family history of psychosis or a personal history of schizotypal personality disorder accompanied by a decline in general functioning (n = 1). At intake, they were also assessed using the Beck Depression Inventory (BDI) and State Trait Anxiety Inventory (STAI) (Table 1).[23, 24] Eighteen ARMS subjects were antipsychotic naïve at the time of scanning, but three subjects were receiving low doses of atypical antipsychotics (risperidone, blonanserin, or aripiprazole) and one was treated with sulpiride. They were also receiving benzodiazepines (n = 3), antidepressants (n = 1), and/or tandospirone (n = 3). The mental condition of each subject was regularly assessed by experienced psychiatrists to check for the emergence of full-blown psychosis at outpatient clinics of the Department of Neuropsychiatry of Toyama University Hospital; five (22.7%) of the ARMS subjects in this group developed schizophrenia fulfilling ICD-10 research criteria[25] and 17 (77.3%) did not develop psychosis during follow up (mean clinical follow-up period after scanning, 15.6 ± 17.4 months).

Table 1. ARMS and FESz subject data vs matched controls (mean ± SD)
ParametersARMS (11 M, 11 F)Controls (11 M, 11 F)Group comparisons
  1. Different typical and atypical antipsychotic dosages are converted into HPD equivalents using the guideline by Toru.[22]
  2. Data missing for one participant. ARMS, at-risk mental state; BDI, Beck Depression Inventory; FESz, first-episode schizophrenia; HPD, haloperidol; SANS, Scale for the Assessment of Negative Symptoms; SAPS, Scale for the Assessment of Positive Symptoms; STAI, State Trait Anxiety Inventory.
Age (years)19.1 ± 4.119.4 ± 4.2F(1,42) = 0.05, P = 0.830
Height (cm)162.2 ± 9.5164.3 ± 8.5F(1,42) = 0.62, P = 0.436
Education (years)11.1 ± 1.613.1 ± 2.6F(1,42) = 8.99, P = 0.005
Parental education (years)13.8 ± 1.712.4 ± 1.6F(1,42) = 7.68, P = 0.008
Medication dose (HPD equiv., mg/day)2.2 ± 3.1 (n = 4)
Duration of medication (months)2.3 ± 4.1 (n = 4)
Time between intake and scan (days)50.8 ± 74.4
Time between scan and onset (months)8.2 ± 9.9 (n = 5)
STAI trait at intake65.3 ± 10.9
STAI state at intake58.4 ± 11.3
BDI at intake24.1 ± 10.0
SAPS total at scanning20.4 ± 10.9
SANS total at scanning 48.5 ± 19.4
 FESz (37 M, 27 F)Controls (37 M, 27 F)Group comparisons
Age (years)24.0 ± 4.725.1 ± 5.0F(1,126) = 1.64, P = 0.203
Height (cm)164.9 ± 7.6167.0 ± 7.5F(1,126) = 2.60, P = 0.109
Education (years)13.5 ± 1.916.5 ± 2.6F(1,126) = 57.55, P < 0.001
Parental education (years)13.0 ± 2.013.2 ± 2.5F(1,124) = 0.50, P = 0.482
Onset age (years)23.1 ± 4.7
Illness duration (months)11.2 ± 12.2
Medication dose (HPD equiv., mg/day)10.3 ± 8.8
Duration of medication (months)8.3 ± 12.6
SAPS total at scanning27.3 ± 21.9
SANS total at scanning53.1 ± 25.2

Sixty-four FESz patients who fulfilled the ICD-10 research criteria,[25] with illness duration ≤1 year (n = 48) or under first psychiatric hospitalization (n = 16) at the time of scanning,[26-29] were recruited from inpatient and outpatient clinics of the Department of Neuropsychiatry of Toyama University Hospital (Table 1). The diagnosis of schizophrenia was confirmed for all patients at least 6 months after illness onset based on information obtained from a detailed chart review as well as their clinical symptoms rated at the time of scanning. They were also screened for other neuropsychiatric conditions (e.g. depressive/manic symptoms) by experienced psychiatrists. All but two of the patients were on antipsychotic medication; 18 were treated with typical antipsychotics, 43 were receiving atypical antipsychotics and one received both typical and atypical antipsychotics.

The control subjects consisted of 86 healthy volunteers recruited from the community, hospital staff, and university students. Given the sexual dimorphism (male < female) and age-related atrophy of the pituitary gland,[30-32] the control subjects comprised two groups that were age- and gender-matched for ARMS (n = 22) and for FESz (n = 64), respectively (Table 1). Although the controls did not receive a full diagnostic interview, they were given a questionnaire consisting of 15 items concerning their personal (13 items; e.g. a history of obstetric complications, substantial head injury, seizures, neurological or psychiatric diseases, impaired thyroid function, hypertension, diabetes, and substance use) and family (two items) histories of illness.[33] They did not have any personal or family history of psychiatric illness among their first-degree relatives.

All subjects in this study (ARMS, FESz, and controls) were screened using the same exclusion criteria (except family history of psychiatric illness, which was applied only to controls). They were right-handed and physically healthy at the time of the study, and none had a history of serious head trauma, severe obstetric complications, neurological illness, substance abuse disorder, or serious medical disease (e.g. impaired thyroid function, hypertension, and diabetes). The FESz and ARMS participants were screened for these conditions using a detailed chart review at scanning (FESz) or direct interview at study intake (ARMS). None of the participants was pregnant or taking exogenous estrogens at the time of the study, but hormone levels as well as menstrual cycle in female subjects were not assessed in this study. All participants were also screened for gross brain abnormalities by neuroradiologists.

The clinical symptoms of the ARMS and FESz subjects were rated at the time of scanning using the Scale for the Assessment of Negative Symptoms and the Scale for the Assessment of Positive Symptoms (SANS/SAPS).[34] Of the 172 participants in this study, 60 controls (35 male) and 37 schizophrenia patients (21 male) were also included in our previous pituitary study.[35] This study was approved by the Committee on Medical Ethics of Toyama University. After a complete description of the study was provided, written informed consent was obtained from all subjects.

Magnetic resonance imaging procedures

The subjects were scanned on a 1.5-T Magnetom Vision (Siemens Medical System, Erlangen, Germany) with a 3-D gradient-echo sequence fast low-angle shots (FLASH) yielding 160–180 contiguous T1-weighted slices of 1.0-mm thickness in the sagittal plane. The imaging parameters were as follows: repetition time, 24 ms; echo time, 5 ms; flip angle, 40°; field of view, 256 mm; and matrix size, 256 × 256 pixels. The voxel size was 1.0 × 1.0 × 1.0 mm. The scanner was calibrated weekly with the same phantom to ensure measurement stability.

To assess the pituitary volume, the images were processed on a Linux PC (Fujitsu, Tokyo, Japan) using Dr. View software (AJS, Tokyo, Japan). Brain images were realigned in three dimensions to standardize for differences in head tilt during image acquisition and were then reconstructed into entire contiguous coronal images of 1-mm thickness perpendicular to the anterior commissure–posterior commissure line. The signal intensity histogram distributions from the T1-weighted images across the whole cerebrum were then used to semi-automatically segment the voxels into brain tissue components and cerebrospinal fluid. The intracranial volume (ICV) was measured to correct for differences in head size as described previously;[36] there were no significant group differences for ICV (ARMS vs their controls, F(1,41) = 0.88, P = 0.353; FES vs their controls, F(1,125) < 0.01, P = 0.984; and FES vs ARMS vs all controls, F(2,168) = 0.43, P = 0.654; Table 2).

Table 2. Intracranial and pituitary volume (mean ± SD)
VariablesARMS (11 M, 11 F)Controls for ARMS (11 M, 11 F)FESz (37 M, 27 F)Controls for FESz (37 M, 27 F)
  1. aSignificantly larger than age- and gender-matched controls. Statistical analysis for the pituitary gland was based on relative volume. Analysis of covariance with age as a covariate and with diagnosis as a between-subject factor was used for the intracranial volume. ARMS, at risk mental state; FESz, first-episode schizophrenia.
Intracranial volume (cm3)1460 ± 1321500 ± 1461500 ± 1471502 ± 150
Pituitary volume (mm3)763 ± 124a697 ± 143802 ± 153a708 ± 140

Pituitary measurements

The pituitary gland volume was manually traced on consecutive 1-mm coronal slices based on a method used by Garner et al.[4] Briefly, we traced around the usually well-defined borders of the anterior and posterior pituitary: the diaphragma sellae, superiorly; the sphenoid sinus, inferiorly; and the cavernous sinuses, bilaterally. As presented in Figure 1, the pituitary stalk was excluded from the tracings, but we included a posterior bright spot, corresponding to the posterior pituitary (the intensity of which is thought to reflect the vasopressin concentration). All measurements were carried out by a trained rater (TT) without knowledge of the subjects' identities or the times of their scans. To determine the reliability of the measurement, a second rater (VL) measured the pituitary volume in a subset of 10 randomly selected brains. Each pituitary volume in these 10 brains was then remeasured after at least 4 weeks by the first rater. Inter- (TT and VL) and intra-rater intraclass correlation coefficients were >0.93.

Figure 1.

(a) Sagittal and (b) coronal views of the pituitary gland manually traced in this study. The pituitary stalk was excluded from the tracings, but a posterior bright spot was included.

Statistical analysis

The relative volume of the pituitary gland ([absolute volume/ICV) × 100]) was analyzed using analysis of covariance (ancova) with age as a covariate and with diagnosis and gender as between-subject factors. The effect of medication type (typical vs atypical for FESz) and outcome (with vs without later transition for ARMS) on relative pituitary volume was also examined on ancova. Post-hoc Scheffé's tests were carried out to follow up any significant main effects or interactions. Spearman's rank correlations were calculated to examine relationships between relative pituitary volume and the clinical variables. Statistical significance was defined as P < 0.05 (two-tailed).


Group comparisons of the pituitary volume

ancovas of the pituitary volume showed significant main effects for diagnosis (ARMS vs their controls, F(1,39) = 4.94, P = 0.032; FESz vs their controls, F(1,123) = 15.58, P < 0.001) and gender (ARMS vs their controls, F(1,39) = 26.39, P < 0.001; FESz vs their controls, F(1,123) = 113.58, P < 0.001) but not diagnosis × gender interaction (ARMS vs their controls, F(1,39) = 0.91, P = 0.346; FESz vs their controls, F(1,123) = 0.66, P = 0.417). Post-hoc analyses showed that both the ARMS (P = 0.030) and FESz (P < 0.001) groups had a larger pituitary volume compared with matched controls, and female subjects had a larger volume than male subjects (P < 0.001; Table 2; Fig. 2). Direct comparison of the pituitary volume between ARMS and FESz showed no significant group difference (F(1,81) = 1.58, P = 0.213).

Figure 2.

Absolute pituitary volume in the at-risk mental state (ARMS) individuals, controls for ARMS, first-episode schizophrenia (FESz) patients, and controls for FESz. Arrows, ARMS individuals with later transition into psychosis. Horizontal lines, mean. Post-hoc test: *P < 0.01, **P = 0.03 (statistical analysis for the pituitary gland was based on relative volume).

These results remained essentially the same even when we added medication dose and duration as covariates, and there was no difference in the pituitary volume between the FESz patients treated with typical (n = 18) and atypical (n = 43) antipsychotics (F(1,58) = 0.30, P = 0.586). Although the ARMS subjects who were taking antipsychotics at scanning (n = 4; pituitary volume, 843 ± 128 mm3) had a larger pituitary volume than antipsychotic-naïve ARMS subjects (n = 18; pituitary volume, 746 ± 120 mm3), the difference was not statistically significant (F(1,19) = 1.73, P = 0.204). The comparison of the pituitary volume between the antipsychotic-naïve ARMS and FESz subjects showed no significant group difference (F(1,77) = 1.60, P = 0.209). When we examined only antipsychotic-naïve ARMS subjects (n = 18) and 18 age- and gender-matched controls, pituitary expansion did not reach significance (F(1,31) = 3.48, P = 0.072). The pituitary volume did not differ significantly between the ARMS subjects who later developed schizophrenia (n = 5; pituitary volume, 803 ± 78 mm3) and those who did not (n = 17; pituitary volume, 752 ± 134 mm3; F(1,19) = 0.87, P = 0.362).

Correlation analysis

The relative pituitary volume did not correlate with age, education, or parental education in all groups. No significant correlation was found between the pituitary volume and the BDI or STAI (state, trait) score in the ARMS subjects. In the FESz group, the pituitary volume was not significantly correlated with onset age or illness duration. In both the ARMS and FESz groups, no significant correlation was found between the pituitary volume and the total scores for the SANS/SAPS or medication (daily dose at scanning, duration of antipsychotic treatment).


This MRI study identified an enlarged pituitary volume in both subjects with ARMS and patients with FESz compared with healthy controls. The effect of medication is an important consideration for pituitary findings,[7, 9, 10, 37] but we found no significant effect of daily dose or duration of antipsychotic treatment on the pituitary volume. Consistent with previous reports,[4, 17] the pituitary volume did not correlate with clinical symptoms in either clinical group. Despite the relatively small number of subjects with ARMS, the present findings suggest that these high-risk subjects may share HPA hyperactivity with FESz patients as a possible indicator of common stress vulnerability.

Pituitary volume in early psychosis

The present finding of enlarged pituitary volume in FESz is consistent with previous MRI studies,[5, 6, 8, 20] supporting the role of HPA hyperactivity in the development of psychosis.[2, 18] To our knowledge, however, there have been only two MRI studies of the pituitary volume in ARMS as compared with healthy controls, which have yielded partly inconsistent results. Garner et al. found no significant group difference in the pituitary volume between the ARMS and controls, but the ARMS subjects who later developed psychosis (ARMS-T) had a larger pituitary volume than those who did not (ARMS-NT).[4] The present results were similar to those of Büschlen et al., who reported that the pituitary volume increased in the order of healthy controls to ARMS-NT to ARMS-T and first-episode psychosis, although the difference between the ARMS-T and -NT was not statistically significant.[20] As discussed by Büschlen et al.,[20] these inconsistencies may be partly due to different ascertainment strategies, as well as different characteristics, of regional psychiatric services. In fact, Garner et al., who included ARMS subjects with a comorbid diagnosis of major depression or borderline personality disorder, suggested the role of the pituitary volume as a predictor of psychotic major depression,[4] whereas the ARMS-T subjects in the present study and those of Büschlen et al.,[20] neither of whom included ARMS subjects with those comorbidities, predominantly developed schizophrenic psychosis. Nevertheless, these MRI studies generally imply that these clinical high-risk subjects could exhibit pituitary expansion at least in some individuals, supporting the notion that an enhanced HPA axis response to stress appears to be part of the biological vulnerability to psychosis.[19] This notion may also be supported by hormone[38-40] and neuroimaging[35] findings in subjects with schizotypal personality disorder (SPD) who have a higher incidence of developing psychosis than the general population,[41] suggesting that distress related to social deficits or incipient psychotic experience could activate the stress response even without florid psychosis.

Possible underlying mechanism of the pituitary expansion

The present structural MRI study could not address the mechanism for pituitary volume changes, but a recent study by Habets et al. showed that higher pituitary volume was associated with increased emotional stress reactivity especially in patients with psychotic disorder.[42] It may be possible that pituitary expansion in the present study reflects HPA axis hyperactivity and a subsequent increase in the size and number of corticotrophs (cells producing adrenocorticotropic hormone; ACTH), which can be explained by an activation of the hormonal stress response.[5, 6] Estrogen treatment, hypothalamic tumor, pregnancy, and primary hypothyroidism also lead to pituitary expansion,[43, 44] but these common causes of pituitary enlargement were excluded in the present subjects.

Antipsychotic medication could also influence HPA activation,[2, 45, 46] but the effect of medication on the pituitary volume remains controversial. Recent MRI studies suggested that atypical antipsychotics might reduce pituitary volume in the course of psychosis,[9, 10, 37] consistent with the notion that antipsychotic medication generally dampens HPA activity in schizophrenia.[1, 2, 46] In contrast, some antipsychotics may increase pituitary volume, possibly by activating prolactin-secreting cells.[5, 7, 47] Although we did not find a direct relation between the pituitary volume and medication (daily dose at scanning, duration of antipsychotic treatment), almost all of the present FESz patients had been taking antipsychotics for a substantial period at the time of scanning (mean, 8.3 months) and significant pituitary expansion of the ARMS subjects diminished when we investigated only antipsychotic-naïve ARMS subjects. Thus, the possibility still exists that the pituitary expansion in the present study was partly related to the effect of antipsychotic medication, which should be further examined.

Pituitary volume and clinical characteristics

In contrast to neuroendocrine observations demonstrating the association of plasma or salivary cortisol levels with prodromal (including depressive and anxiety) or psychotic symptoms in ARMS subjects,[14-17] this and previous MRI studies in ARMS found no significant correlation between the pituitary volume and global psychopathology, general functioning, or psychotic symptomatology.[4, 17] Direct comparison of plasma and MRI findings in ARMS also showed that pituitary volume did not correlate with either plasma cortisol level or number of glucocorticoid receptors.[17] Our previous study, however, identified a significant relationship between ongoing pituitary expansion and treatment response or severity of positive psychotic symptoms in FESz,[8] suggesting that it is longitudinal pituitary changes during early phases of the illness that are relevant to clinical manifestations of psychosis. Interestingly, a recent study of cortisol level emphasized the role of longitudinal HPA changes in the development of psychosis.[18] Thus, further study of the association of longitudinal pituitary volume changes with HPA functioning and clinical characteristics (e.g. symptom severity, later transition into psychosis) is required to examine the potential role of HPA activity in the emergence of psychosis in vulnerable individuals.

Methodological considerations

A few possible methodological considerations in this study should be taken into account. First, the sample size of the present ARMS group (especially those who later developed psychosis) was relatively small and the clinical follow-up period was short for some individuals. Although we found no significant difference in pituitary volume between the ARMS subjects with and without later transition to psychosis, whether the baseline pituitary volume could predict onset of psychosis should be tested in a larger, well-defined high-risk cohort. Second, although the present findings of pituitary enlargement in early psychosis are thought to reflect state-related HPA axis dysregulation, we did not directly assess pituitary function. The pituitary gland is also considered to be sensitive to prolactin-elevating antipsychotics[7, 47] and a recent study reported hyperprolactinemia in antipsychotic-naïve ARMS subjects.[48] The present findings replicated the sexual dimorphism of the pituitary gland volume (female > male),[32] potentially reflecting different endogenous estrogen levels.[49] We did not, however, assess prolactin or estrogen level in this study. Thus, additional assessment of both pituitary volume and hormone levels (e.g. cortisol, ACTH, prolactin, and estrogen) is required. The present study might be also limited by a lack of urine toxicology screening for substance use. Finally, given that HPA axis functioning also appears to be affected in major depressive disorders[50-52] and that Garner et al. found an enlarged pituitary volume prior to the onset of psychotic major depression,[4] further investigation of the disease specificity of pituitary findings is warranted.


Both the ARMS and FESz subjects had significant enlargement of the pituitary gland, presumably reflecting activation of the hormonal stress response during early psychosis. Given that the pituitary gland is a dynamic organ reflecting state-related HPA axis dysregulation, longitudinal study of the pituitary volume and its relation to clinical characteristics, as well as hormone levels, is required to further understand the role of HPA functioning in the emergence of psychosis.


This research was supported in part by Grants-in-Aid for Scientific Research (C) (No. 22591275, 24591699) and Grants-in-Aid for Scientific Research (B) (No. 24390281) from the Japanese Society for the Promotion of Science, Health and Labour Sciences Research Grants (Comprehensive Research on Disability, Health and Welfare, H23-Seishin-Ippan-002 and H23-Seishin-Ippan-009), and a Research Grant from the JSPS Asian Core Program. The authors are grateful to Ms Valentina Lorenzetti for assistance with MRI analysis. The authors have no conflicts of interest to declare.