Introduction

Currently, the ultra-high risk (UHR) criteria, especially the attenuated psychotic symptom (APS) syndrome, are mainly used for the early detection of psychosis; pooled conversion rates in UHR samples have risen to 37% over more than 4 years [1]. UHR criteria were developed on predominately adult samples and only later applied to children and adolescents (CAD). However, this was done without the consideration of possible developmental aspects. Thus, unsurprisingly, pooled conversion rates in clinical UHR samples of CAD were significantly lower than in adult or mixed-age samples [1], and within a clinical APS-syndrome sample, significantly less 12–14-year-olds than 15–22-year-olds developed psychosis [2].

In community samples of children, high prevalence rates of mainly non-persistent hallucinatory experiences of little clinical relevance were reported [3]. Furthermore, compared to 16–40-year-olds, 8–15-year-olds of the Swiss general population reported higher frequencies of perceptual APS and lesser clinical significance of non-perceptual APS [4]. We, therefore, examined if a similar age-effect can be detected in a clinical never-psychotic sample of the same ag-range referred to a specialized service for clinical suspicion of developing psychosis.

Methods

Sample

Excluding 42 patients with past or present psychosis, data from 133 patients who were examined in the Bern Early Detection and Intervention Center for Mental Crises (FETZ Bern) for a beginning psychosis between November 2009 and April 2015 was analyzed (Table 1). As required by the local ethics committee, patients, and if minors—their legal guardians—gave their informed consent for their anonymized clinical data to be used in scientific analyses and publications.

Table 1 Sample characteristics (N = 133)
Full size table

Assessments

APS and brief intermittent psychotic symptoms (BIPS) and the fulfillment of onset/worsening and frequency requirements in UHR criteria (see supplementary material 1) were assessed by items P1–3 and P5 (non-perceptual), and P4 (perceptual) of the Structured Interview for Psychosis-Risk Syndromes [5] that also includes a genetic risk and deterioration syndrome (GRDS) (see supplementary material 2 for case examples). Proxy measures of clinical relevance, i.e., current axis-I disorders and/or functional deficit, were assessed with the Mini-International Neuropsychiatric Interview [6, 7] and the Social and Occupational Functioning Assessment Scale (SOFAS; score ≤70 and ≤60) [8].

Data analyses

Using SPSS 24, binary logistic regression analyses using “enter” were performed to assess the effects of different age groups (Table 2) on UHR symptoms and their criteria requirements. The age group with a peak in the onset of first-episode psychosis (20–24 years) served as the reference group. Logistic regression analyses were also used to assess the effects of UHR requirements and their interaction with age on low functioning, and on the presence of any axis-I disorder. The interaction with age was considered as relevant when both backward and forward logistic regression analyses equally selected the interaction term as a predictor.

Table 2 Age effect on presence of APS and/or BIPS
Full size table

Results

Age effect on prevalence of ultra-high risk symptoms

Eighty-five (64%) patients reported APS (n = 81; 61%) or BIPS (n = 10; 7%). Thirty-three (25%) met the APS syndrome, two (2%) the BIPS syndrome, and none the GRDS. Any perceptual APS/BIPS (P4) was reported by 57 (43%) and any non-perceptual APS/BIPS by 59 (44%) patients. The most frequent non-perceptual APS/BIPS were unusual thought contents (P1; n = 39, 29%) and persecutory ideas (P2; n = 35, 26%), while grandiose ideas (P3; n = 4, 3%) and disorganized communication (P5; n = 7, 5%) were rare.

In correspondence to the results in the community sample [4], perceptual APS/BIPS were significantly more frequent in younger age groups below the age of 16 (Table 2). This was mainly due to higher prevalence rates of BIPS (β = 3.03, SD = 1.17, Wald test: \(\chi_{(1)}^{2}\) = 6.67, p < 0.01, odds ratio (OR) = 20.71, 95% confidence interval (CI): 2.08/206.65), in particular hallucinations (β = 2.67, SD = 1.19, Wald test: \(\chi_{(1)}^{2}\) = 5.07, p < 0.05, OR = 14.50, 95% CI: 1.42/148.57), in the 8–12-year-old group.

Clinical relevance of ultra-high risk symptoms

Contrary to the strong association of mental disorder and functional deficit with APS reported in the community study [4], no association between APS/BIPS or APS requirements with current axis-I disorder or functional deficit was revealed in univariate analyses.

Age effect on clinical relevance of ultra-high risk symptoms

In line with the community study [4], analyses of the effect of the interactions between age and presence of APS, as well as of their onset and frequency requirements on the presence of a mental disorder only revealed a significant, robust interaction effect on mental disorders between age and APS meeting onset requirements (goodness-of-fit: \(\chi_{(1)}^{2}\) = 4.21, p < 0.05; β = 0.06, SD = 0.03, Wald test: \(\chi_{(1)}^{2}\) = 4.86, p < 0.05, OR = 1.07, 95% CI: 1.01/1.13).

Contrary to the community study [4], however, in that interaction terms between age and both APS and APS-onset requirement, had a significant effect on presence of a functional deficit, no effect of interaction terms on the presence of a functional deficit was revealed in the clinical sample.

Discussion

UHR criteria were developed in predominately adult samples, and accumulating evidence indicates that they might not be applicable to CAD samples without sufficient consideration of developmental peculiarities [1,2,3,4, 9]. Not only might UHR criteria, essentially the APS syndrome, be less predictive of psychosis in CAD [1, 2] but UHR symptoms, primarily perception-related phenomena, might also be more frequent [3, 4] in clinical and non-clinical samples of less-than-16-year-olds. Furthermore, unusual, paranoid and grandiose thought contents might be less associated with indicators of clinical relevance [4], in particular, in young non-clinical samples. In our clinical sample, however, the age-effect of UHR symptoms on clinical relevance in terms of a functional deficit was less obvious, indicating that other symptoms and disorders—if already frequently present as in our sample—might impact functioning more severely than APS.

Taken together, the present and earlier studies [2, 4] point towards an age threshold around the transition from early-to-late adolescence, i.e., around age 15/16 years. This age threshold is likely related to the maturation processes of major cognitive abilities such as source monitoring, deductive reasoning and attributional style [4] that are in the focus of cognitive-behavioral therapy techniques used in predominately adult UHR samples, including normalization, cognitive and behavioral experiments, and cognitive restructuring to improve stress- and symptom-management [10]. Thus, the replicated age-effect on perceptual APS in this clinical sample highlights the need to examine ways to distinguish clinically relevant from intermittent hallucinatory phenomena of little-to-no clinical relevance. Until then, the advice of the EPA guidance on the detection of clinical high-risk states [1] should be followed, especially in under-16-year-olds: “… CHR criteria should only be used and communicated with outmost care in children and young adolescents in whom they should nevertheless be assessed and monitored.”