Treatment Approaches for First Episode and Early-Phase Schizophrenia in Adolescents and Young Adults: A Delphi Consensus Report from Europe

Introduction

Sharing many overlapping characteristics with adult-onset schizophrenia, first-episode psychosis (FEP)/early-onset schizophrenia (EOS), ie, onset prior to 18 years of age, is a relatively rare form of schizophrenia.^1–9 According to a recent meta-analysis, 2% of people with schizophrenia had their onset prior to age 14 (schizophrenia-spectrum disorders = 3%) and 8% prior to age 18 (schizophrenia-spectrum disorders = 12%).¹⁰ Furthermore, compared to adult-onset schizophrenia, EOS is characterised by a longer duration of untreated psychosis, more pronounced negative and cognitive symptoms since onset, greater long-term disability and less favourable prognosis.^1–3,11,12

Compared to adult-onset schizophrenia, the diagnostic process is complex given the preceding psychopathology (also termed the clinical high-risk stage),^13,14 many ongoing physiological changes and peak time emergence of other mental disorders during this developmental age,¹⁰ when disruption of early neurodevelopmental patterns can occur, particularly during the prenatal/perinatal period and adolescence.^15–21 Genetic susceptibility enhanced by environmental or chemical factors and traumatic events appears to trigger an abnormal development of some brain regions, resulting in alteration of white matter, greater loss of grey matter, excess synaptic pruning, and dysregulation of stress signalling pathways.^16–19,22 The differentiation of impairments in brain connectivity depends on the developmental period during which these triggers act, influencing disease onset and trajectory. Many of the implicated risk factors²³ and symptoms peak between the age of 13–18 years²⁴ and may overlap with those of other disorders, such as autism spectrum and other conditions,^16,17,25,26 pointing to schizophrenia as a cluster of disease entities with multiple causes.^26,27

Accordingly, the management of FEP/EOS currently includes multimodal approaches integrating pharmacological and nonpharmacological interventions (psychosocial, cognitive-behavioural therapy, etc.),^28,29 which can improve the disease course and long-term outcomes compared to either of these approaches alone.^2,4,28,29 Pharmacological interventions available for EOS mostly rely on the same antipsychotic agents used in adult-onset schizophrenia.^11,30–33 Antipsychotics primarily consist of dopamine D2 receptor antagonists (first-generation antipsychotics, FGAs) and serotonin-dopamine antagonists/partial agonists (second-generation antipsychotics, SGAs). Different effects on these and other receptors are also responsible for different side effects, including weight gain, lipid abnormalities, glucose abnormalities, prolactin elevation, sedation, extrapyramidal effects and akathisia.^30,34

SGAs approved in Europe for the treatment of schizophrenia include aripiprazole, asenapine, brexpiprazole, cariprazine, lurasidone, olanzapine, paliperidone, quetiapine, risperidone, and ziprasidone, as well as clozapine for treatment-resistant schizophrenia.³⁵ Of these SGAs, only three have specific European Medicine Agency (EMA) approval for use before 18 years of age: aripiprazole and paliperidone approved for people aged 15–17 years old, and lurasidone approved for use at 13–17 years of age.^11,30–32 To date, in adults and youth, there is no direct evidence of higher efficacy of any specific antipsychotic over another, except for clozapine.^{31–33,36,37} A recent network meta-analysis,³³ which specifically addressed the use of non-clozapine antipsychotics in adolescents with EOS, concluded that all agents evaluated, with the exception of ziprasidone and asenapine, had comparable effectiveness in terms of PANSS total symptom change and other outcomes and that adverse reaction/safety profiles were consistent with the findings in adult patients. The study, however, did not include the antipsychotic lurasidone, which has recently been approved by EMA for treatment from the age of 13 years old.³¹ The data available to date demonstrate lurasidone similar efficacy to other oral SGAs but with less weight gain and a lower risk of all-cause discontinuation.³¹

In reference to this latter aspect, children and adolescents seem more susceptible to acute extrapyramidal side effects, sedation, withdrawal dyskinesia, hyperprolactinemia, weight gain and metabolic abnormalities than adults.^30,38,39 Thus, in principle, the choice of an agent should be driven by its tolerability profile.³

In any case, because adolescence is a critical stage of development, early interventions can increase the possibility of recovering and pursuing psychological wellbeing. Specifically, because EOS dramatically interferes with age-related developmental milestones, such as independent living, vocational training and social integration, early intervention is paramount in preventing cumulative disability for the patient as well as an escalation of costs for society.^11,16,30,38

To date, recommendations for the management of EOS mirror those for adults and are quite heterogeneous, with differences stemming from local health care policies, organization of medical services, treatment schools, economic resources, and cultural settings.^40,41

The aim of this work was to explore any critical areas in the management of FEP/EOS by means of a computer-based survey involving European specialists with expertise in paediatric schizophrenia to provide guidance to clinicians on the best, evidence-based management strategies for EOS.

Materials and Methods

Study Design and Scope

We employed the Delphi methodology, which is widely used as an exploratory tool to gain knowledge in areas that are difficult to cover by means of evidence-based strategies.^42,43 In general, the process begins with an open-ended research question being discussed among a group of content experts (Steering Committee) through an iterative process involving the sharing of opinions, professional experience, and scientific evidence. The process (question sourcing) leads to the development of a list of statements/items (Delphi questionnaire), which is then submitted to a broader panel of professionals (panellists) to survey their level of agreement on the topics proposed (Delphi rounds of consensus).^42,43 The process is completed when feedback converges, providing no new elements of insight (reaching saturation).

The present study, which took place from November 2020 to May 2021, addressed the unmet needs in the management of FEP/EOS, asking which were the areas needing priority attention. The Steering Committee was composed of 5 adult psychiatrists with a focus on first-episode and young adults with psychotic disorders, 1 child and adolescent psychiatrist and 2 experts in both child/adolescent and adult psychiatry. Professionals invited to discuss the topic were chosen among a group of internationally recognized experts in the treatment of child and adult psychiatry and/or schizophrenia with experience in guidelines committees and clinical research who were actively practicing in Europe.

Development of the Delphi Questionnaire

This first step of question sourcing served to collect an initial pool of feedback on areas/domains that the experts considered most critical based on their clinical practice. Comments were then analysed and coded into themes, grouping subsets of related items of enquiry. The organization of themes and items into a coherent and meaningful set of statements for the Delphi questionnaire was further informed by a literature review on the management of FEP/EOS in European countries (reviews, trials, guidelines). The first draft of the Delphi Questionnaire was submitted to the members of the Steering Committee for critical appraisal and improvement of the draft to ensure that items were relevant to the research question, were clearly worded and did not overlap with previous items.

The second draft was then submitted for validation to a restricted group of colleagues. Once the final version of the questionnaire was approved by the Steering Committee, the document was placed on a dedicated online platform and submitted to a list of participants. The candidate panelists were selected by members of the Steering Committee who provided contacts of clinical specialists in the field of child and adolescent and adult psychiatry (psychiatrists, psychologist, neuropsychiatrists) practicing in Europe with ≥4 years of clinical experience in the treatment of FEP/EOS and who considered themselves “expert clinicians” in the use of antipsychotics for this indication.

Delphi Rounds of Approval

Participants were invited to participate in the survey via e-mail. Replies from the first round were collected in an anonymous fashion. Levels of agreement/disagreement for each statement were captured by a 5-point Likert scale (1= strongly disagree, 2= disagree, 3=agree, 4= more than agree, 5= strongly agree). Consensus was predefined as >66% of the panel agreeing/disagreeing with any given item.

All responses from the first round were reviewed by the Steering Committee and commented on during a virtual roundtable discussion. Any items that did not yield a clear-cut agreement, were in contrast with the current scientific evidence (statement 18.4), or generated new areas of enquiry (statement 2.7, 18.1, 18.2, 18.3, 18.5) were discussed by the Steering Committee in an iterative process. To gain insight into the opinions and experience of the panellists about these themes with unclear results, another set of statements was sent out to the participants for a second round of feedback. In the invitation to the second Delphi round, panellists received an overview of the results from the first round and a selection of recent evidence on the topics addressed by the statements included in the second round according to standard Delphi methodology.^42,43

The study was based on a survey that does not involve the participation of human subjects nor patient data management. Consequently, this study did not require ethical approval. All experts involved in the Delphi survey were informed of the study’s objectives and the possibility of publishing the results in a peer-reviewed article. The participation was voluntary. The panelists expressed their consent to participate in the survey after logging into the secure online survey platform by actively clicking on the appropriate consent box.

Results

The Delphi process involved two rounds of questioning to reach consensus (Figure 1). The questionnaire for Round I covered 5 main topics/domains spanning the patient pathway from diagnosis to maintenance treatment, namely: timely detection of signs of FEP/EOS and formal diagnosis, access to the care of individuals with symptoms and signs of FEP or EOS, treatments and outcomes, nonpharmacological treatments, strategic management, and clinical practice/scientific evidence. Overall, the questionnaire featured 18 main statements, each developed into one or more items for a total of 89 items (see Appendix 1). All items asked panellists for their agreement on what they believed should be standard practice in their culture/treatment setting.

Figure 1 Delphi study flowchart.

Of 92 professionals contacted to participate in the survey, 10 contacts were excluded due to e-mail delivery issues, and 9 questionnaires were incomplete upon the closing date and, therefore, discarded. For the first round, complete responses were gathered from 54 mental health practitioners (74% of all professionals contacted), equally balanced between males (46%) and females (54%), from eight different European countries (1 from Bosnia and Herzegovina, 2 from Denmark, 13 from Germany, 12 from Italy, 1 from Montenegro, 7 from Serbia, 10 from Spain, 7 from UK). In the second round, responses were gathered from all but 6 respondents from the first round.

Among the 90 items, a consensus was reached for 83 (92%) items (80 with a positive consensus, 3 with a negative consensus, ie, 7.1, 18.6, 18.8), while only 7 items (2.7, 4.4, 7.3, 7.5, 12.2, 18.1, 18.4) failed to reach the >66% threshold for establishing consensus. Among the 80 items that gained positive consensus, the agreement was strong (scores 4 and 5 more frequent than 3) for 59 (74%) statements (Figure 2). On the other hand, among the 3 items that gained a negative consensus, the agreement was strong (score 1 more frequent than 2) for none.

Figure 2 Continued.

Figure 2 Continued.

Figure 2 Continued.

Figure 2 Continued.

Figure 2 Continued.

Figure 2 (A) Delphi round I levels of agreement and distribution of responses – Topic 1. (B) Delphi round I levels of agreement and distribution of responses – Topic 2. (C) Delphi round I levels of agreement and distribution of responses – Topic 3. (D) Delphi round I levels of agreement and distribution of responses – Topic 4. (E) Delphi round I levels of agreement and distribution of responses – Topic 5. (F) Delphi round I levels of agreement and distribution of responses – Topic 6.

Based on the responses from the first round, the Steering Committee decided that 3 of the 18 main statements required rerating (Figure 1), leading to 4 new items (19.1–19.4) and 7 edited items (S. 2.7 and 18.1–18.6). Aside from one question on the diagnostic process (item 2.7), all other items required further insight concerning treatment aspects. Of the 11 items submitted to the panellists for Round II, consensus (either positive or negative) was reached for 7 items (64%), leaving 4 statements with no strong consensus position (ie, 18.2, 18.3, 19.1, 19.2) (Figure 3).

Figure 3 Topic 1 – Timely detection of signs of psychosis/EOS and formal diagnosis. Topic 6 – Clinical practice/scientific evidence.

Timely Detection and Access to Care

In general, panellists broadly acknowledged the need for repeated evaluations to reach the diagnosis of FEP/EOS (S. 1) and the benefit of rapid service engagement towards better future outcomes (S. 3).

Regarding the diagnostic tests to be considered gold standards in a differential diagnosis (S. 2.1–2.11), preferences were given (in order from highest to lowest agreement) to history taking, cognitive assessment, history of substance abuse, thorough physical assessment and laboratory work, urine toxicology, diagnosis of EOS with validated tools, using magnetic resonance brain imaging, and electroencephalography (70%). Conversely, lumbar puncture (S. 2.7) was not considered the gold standard test (only 35% in favour in the first round), unless there was clinical evidence of neurological abnormality (85% agreed in the second round).

As to obstacles to timely detection (S. 4.1–4.11), panellists mainly attributed this to patients and/or family members and their lack of awareness of mental health issues or the need for medical referral. Many (81%) also attributed the delay to be linked to missed recognition by health care professionals (S. 4.8). The lack of affordable services (S. 4.4) received mixed responses without a consensus.

Increasing education/awareness through mental health literacy campaigns (S. 5) was considered appropriate for all groups, from general practitioners and health professionals to the general population.

Treatments and Outcomes

Panellists strongly agreed that young patients should receive both pharmacological and specific evidence-based psychosocial interventions (S. 6) and be involved in adherence-enhancing strategies (S. 11). Despite strong agreement on the higher sensitivity of younger patients to medication adverse effects (S. 7.2) there was mixed feedback regarding whether young patients required lower antipsychotic doses than adults (S. 7.3). Likewise, most panellists tended to disagree (without reaching a clear-cut consensus) that adolescents/young adults should only be treated with antipsychotics that are licenced for that age group (S. 7.5.).

Regarding the need for periodic evaluations of side effects (S. 9) and metabolic parameters throughout pharmacological treatment (S. 7.6), there was broad agreement on the timing of evaluations. However, 13% did not agree on an ECG evaluation upon reaching the steady dose of antipsychotics.

Nonpharmacological Interventions

Among the most frequently employed nonpharmacological interventions, almost all interventions listed (S. 12.1 to 12.11) received broad agreement from panellists, with the preference (in order of level of agreement achieved) for social training skills and supported employment (100%), targeting negative and cognitive symptoms and physical activity (98%), occupational therapy (91%), assertive outreach (87%), and art/music therapy. Psychodynamic therapy was the only intervention not considered useful (negative consensus).

While panellists agreed that intervention services for adolescents/young adults should be connected into an integrated network and on having dedicated transitional age services (S. 15), no clear recommendation emerged regarding the duration of early intervention services (S. 14.1–14.5), although most agreed on a timeframe between 2 and 5 years.

Clinical Practice/Scientific Evidence

Regarding the goals of antipsychotic treatment (S. 16.1–16.6), none emerged as clearly more important than the others, with all being ranked as “the most important”.

Regarding treatment strategies adopted in cases of nonresponse and on how long to wait after treatment failure (S. 18.1–18.3), most panellists agreed on waiting 4 weeks for a response, with some even waiting 6 weeks before considering a change in therapy, whereas the 2-week waiting period (S. 18.1) did not yield a consensus either in favour or against. Therefore, the statement was further explored in Round II with the items edited (Figure 3), splitting the waiting period into one, two, three, four and six weeks. In Round II, the vast majority (94%) of panellists shifted opinions towards waiting at least one week after nonresponse before switching antipsychotics.

Panellists were markedly more inclined to switch to another antipsychotic monotherapy (S. 18.5) than adding a second antipsychotic (S. 18.6). Regarding the switch to clozapine (S.18.8–18.10), very few (15%) agreed to switch after one treatment failure, while most were in favour of switching to clozapine after two and three failures (70% and 80%, respectively). However, 20% of respondents did not agree to the switch even after three treatment failures. Because such a response is in contrast with evidence from the literature, the topic was further explored in the second round by asking which actions the panellists would take instead of clozapine in case of treatment failure. This was explored in Round II through one new statement divided into 4 items (S. 19.1–19.4, Figure 3). While 79% agreed on adding a psychosocial or psychotherapeutic intervention, no consensus was reached for the proposals of combining two antipsychotics or adding a mood stabilizer, whereas most agreed on not adding ECT (77% negative consensus).

Finally, as to dosages used in case of suboptimal/nonresponse (S. 18.4), more than half (65%) disagreed on increasing the dosage above the package insert maximum level, a percentage that increased by 3% when the question was resubmitted to panellists upon round II.

Tables 1–3 summarize the main scientific evidence concerning the issues above, with the intent of providing some reference to specialists in their decision-making.

Table 1 Treatment Strategies in Case of Nonresponse to an Antipsychotic Treatment and Evidence from the Literature on Their Appropriateness in EOS

Table 2 Diagnostic Tools That are Not Considered Standard Practice and Evidence from the Literature on Their Role Within EOS

Table 3 Nonpharmacological Treatments Receiving Mixed Feedback and Evidence from the Literature on Their Role Within EOS

Discussion

We aimed to survey the opinions of European specialists on the main aspects of the management of patients with FEP/EOS and to explore areas showing a need for some clarification. Overall, panellists’ feedback indicated that there is broad agreement on diagnostic standards and on the potential of multimodal approaches, similar to data in adults.³⁸ However, opinions varied more widely on pharmacological strategies, treatments to use in cases of failure, and adequate dosing for younger patients. In particular, the responses suggested a careful pharmacological approach, with particular attention to safety aspects.

The mixed responses received on the waiting time before switching to another monotherapy (S. 18) were consistent with many different ranges (from 2–8 weeks) indicated in guidelines⁴⁴ and reflected the lack of uniform definitions for the concepts of adequate treatment-response time, nonresponse, treatment resistance and response to suboptimal treatment across clinical trials.⁴⁴ However, a diagnostic test meta-analysis⁴⁵ performed on individual patient data showed that patients without even minimal improvement by the second week of antipsychotic treatment are unlikely to improve thereafter. Nevertheless, few data for EOS and first-episode schizophrenia patients are available, and the results are driven by multiepisode adults with schizophrenia. As further studies explore this timeframe, it is also noteworthy that the Treatment Response and Resistance in Psychosis (TRRIP) Working Group encouraged a more consistent use of the term treatment resistance, ie, referring exclusively to the persistence of significant symptoms, despite adequate treatment in at least two antipsychotic treatment trials lasting 6 weeks each, and during which there was no appreciable symptomatic or functional improvement, as measured by validated rating instruments.

Alternative strategies in the case of nonresponse commonly include a switch to another nonclozapine molecule, antipsychotic dose increase, combination therapy, which generally lacks high-quality evidence,^46–49 and a switch to clozapine, which has the only clear evidence for efficacy in treatment-resistant schizophrenia.^50,51

Regarding increasing antipsychotic dosages above the maximum recommended dose in the event of inadequate treatment response, this was addressed by two studies, a meta-analysis and a Cochrane review,^46,52 which found no additional benefit from dose escalation in cases of initial nonresponse to standard-dose therapy.

More recently, a dose–response meta-analysis investigated the minimum effective dose (MED) and near-maximal and effect size for 20 second-generation antipsychotics.⁵³ The study demonstrated that the efficacy of antipsychotics increased proportionally with doses up to the risperidone equivalent of 3–5 mg/dL and then plateaued. For antipsychotics following an increasing dose–response curve (vs a bell curve), the authors hypothesized that higher-than-licenced doses could be more efficacious.⁵³ A subsequent study⁵⁴ of the relationship between 2-fold or 3-fold MED doses and specific side effects confirmed that a dose increase of antipsychotics led to greater differences in terms of side effects rather than efficacy. The authors concluded that while antipsychotic doses can be increased 2- and 3-fold of the MED for increased efficacy, patients should be closely monitored for the emergence of side effects.

Combination therapy is estimated to be quite common in daily clinical practice, occurring in 10–30% of cases.⁵⁵ However, to date, there is no high-quality evidence from blinded randomized trials to support the benefits of combined therapy,⁵¹ despite some uncontrolled real-world evidence.^56,57 Because antipsychotic polypharmacy raises concerns about drug-drug interactions and increased adverse reactions, guidelines restrict this strategy to cases in which other treatments, including clozapine, have failed or are impractical, discouraging its use for extended periods of time.^2,55

Interestingly, while there is little evidence to support the respondents’ preference for antipsychotic polypharmacy, there are several lines of evidence to support the use of clozapine.^5,33,58–63 A network meta-analysis³² comparing antipsychotic efficacy and tolerability in children and adolescents measured changes in overall symptoms of schizophrenia, positive and negative symptoms, response, dropouts, quality of life, social functioning and side effects and found clozapine to be significantly more effective than all other analysed antipsychotics, although it was associated with significant weight gain. Clozapine’s superiority among antipsychotics was also observed in adults (except for negative symptoms),³³ independent of specific schizophrenia subtypes and degree of treatment resistance. Nonetheless, clozapine appears underused, and the switch to clozapine in our survey gained low support among panellists (20% would recommend alternative routes even after three treatment failures), likely reflecting the concern about relevant adverse effects with clozapine,⁶⁴ including haematological adverse events (HAEs) and the need for systematic haematological monitoring.¹ However, HAEs have been relatively low. A study on clozapine-related HAEs in children and adolescents found 13% of the cases were neutropenia and 0.6% were agranulocytosis.⁶⁵ These rates are consistent with a subsequent study finding higher rates of mild and moderate neutropenia compared to adults (the majority of which resolved favourably), without cases of agranulocytosis or severe infection.⁶⁶

A major limitation of the study should be acknowledged, that is the small number of expert participants involved. However, to minimize the potential risk for selection bias, panelists were identified according to their long-term experience in the field and on European distribution.

The panel believes that targeted clinical trials are needed to specifically test therapeutic algorithms and the timing of changing medications to treat resistant EOP. Moreover, population-based studies can help provide clinically useful information about the medium- and long-term safety profile of different medications and combinations of medications used under real-world conditions.

Conclusion

Taken together, responses from panellists reflected a robust consensus of relevant issues in youth and young adults with psychosis/EOS during the vulnerable transitional age regarding diagnostic clues and comprehensive, integrated management. However, uncertainty exists in adapting adult pharmacological approaches to younger age groups while being alert to higher adverse-effect sensitivity of the first-episode/EOS population. Indeed, safety arguments seem to play a major role in the treatment decision-making process. Targeted clinical trials and systematic dissemination across Europe of current scientific evidence on the value of early intervention services also summarized in this Delphi consensus report is hoped to contribute to standardized and improved quality of care for patients with early-phase psychosis and schizophrenia.