talk with our admissions team to get started.
Contact Us
Pand Health

Psychosis and Regression in Down Syndrome

A diagnostic framework for adolescents and young adults, roughly ages 10 to 30, presenting with new psychotic symptoms, catatonia, or functional loss.

Scope and framing

What this review covers

This review addresses new-onset psychotic symptoms, catatonia, and functional regression in individuals with Down syndrome (DS), with emphasis on separating Down Syndrome Regression Disorder (DSRD) from primary psychotic disorders, Alzheimer's disease-associated dementia, catatonia, delirium, and reversible medical mimics. Treatment is addressed insofar as differential response to therapy carries diagnostic information, a critical point, since the response profile (immunotherapy versus benzodiazepine versus antipsychotic) is one of the few discriminating tools available.

The central practical message: in this age band, apparent psychosis in Down syndrome is far more often a manifestation of DSRD, depression with psychotic features, catatonia, or a reversible medical condition than a primary schizophrenia spectrum disorder, and it is essentially never Alzheimer's dementia.

Terminology note: "Down syndrome regression disorder" is the consensus term. Earlier literature uses "Down syndrome disintegrative disorder" (DSDD), "unexplained regression in Down syndrome," and, in the 1946 index description, "catatonic psychosis" (Rosso, Fremion, Santoro et al., 2020).

Background

Why the diagnostic base rates differ in Down syndrome

Intellectual disability (ID) as a class carries roughly 3 to 4 times the general-population rate of psychotic disorders, but Down syndrome appears to be a relative outlier within ID.

378 vs. 371

In the Leicestershire population comparison, the non-DS intellectual disability group was significantly more likely to carry a diagnosis of schizophrenia or paranoid state, while the DS group more often had depression and dementia (Collacott et al., 1992).

n = 245,749

The Western Australian register linkage found schizophrenia greatly over-represented in ID overall (3.7 to 5.2% by birth cohort), yet DS distinctly under-represented among dual-diagnosis cases (Morgan et al., 2008).

SPR 0.6

A Scottish cohort of 186 adults with DS found a standardized prevalence ratio for mental ill-health of 0.6 (95% CI 0.4 to 0.8) versus all-cause ID, falling to 0.4 when organic disorders were excluded (Mantry et al., 2008).

32% vs. 74%

A meta-analysis across neurogenetic syndromes found the lowest pooled prevalence of above-threshold psychiatric symptoms in DS (32%) versus 74% in Prader-Willi syndrome. For contrast, 22q11.2 deletion carries roughly 30% lifetime schizophrenia risk.

Clinical corollary: a diagnosis of schizophrenia in a 19-year-old with Down syndrome should be regarded as a low-prior-probability hypothesis requiring active justification, not a default label for hallucinatory or disorganized behavior.

The first discrimination

Is this actually psychosis?

Before differentiating causes of psychosis, self-talk and imaginary companions must be excluded as normative phenomena. Imaginary companions occur in 28 to 65% of typically developing children aged 5 to 12; where developmental age lags chronological age substantially, these phenomena persist well into adolescence and adulthood. The Developmental Behavior Checklist codes "talks to self or imaginary people or objects" as a communication-disturbance item rather than a psychosis item. The diagnostic signal is change in character plus functional cost, not the presence of self-directed speech.

FeatureBenign self-talk / imaginary companionConcerning for psychosis or regression
ChronologyContinuous from early childhood (Hua, 2021).New onset, or a distinct change in character (Prout et al., 2025).
ContentNeutral, playful, social rehearsal (Einfeld et al., 2006).Threatening, persecutory, frightening (Prout et al., 2025).
RedirectabilityInterruptible; the person re-engages.Preoccupation displaces preferred activities.
Functional costNone.Decline in ADLs, sleep, self-care, school.
AccompanimentsNone.Skill regression, self-injury, insomnia.
Stress relationMay increase but stays manageable (Fotheringham & Thompson, 1994).Escalates and becomes uncontrollable.

Instrument selection

No psychosis-specific instrument has been validated in Down syndrome. The PANSS and PSYRATS depend heavily on introspective self-report and perform poorly in intellectual disability; the core methodological problem is separating chronic neurodevelopmental features from acute psychotic change, which requires baseline-referenced observation rather than cross-sectional rating. The PAS-ADD detected 71% of cases when two or more core symptoms were active, but detection rate correlated positively with IQ, meaning sensitivity falls precisely where it is most needed. The Global Down Syndrome Foundation adult guideline recommends DSM-5 criteria supplemented by the DM-ID-2, which substitutes observable behavioral equivalents for self-report items.

The leading hypothesis

Down Syndrome Regression Disorder

DSRD is the single most important diagnostic consideration for subacute psychosis or functional loss in this age band. It is characterized by subacute loss of previously acquired adaptive, cognitive, and social functioning, comprising mood lability, reduced participation in activities of daily living, new-onset insomnia, social withdrawal, autistic-like regression, mutism, and catatonia, with an acute phase followed by a chronic phase in which baseline may not return. The 2022 international Delphi consensus fixed the nomenclature at 78% agreement and reached 96% agreement on diagnostic criteria.

Age and tempo

Mean age at regression was 17.5 years in the 35-case international database, 11.4 years in the Worley cohort, and 20 to 24 years in the French series. The NEJM review cites mean onset at 11 to 14 years. The largest multicenter cohort (n = 164) enrolled ages 10 to 30.

Psychosocial triggers

Regression cases had six times as many identifiable stressors as matched controls. In one series all 14 individuals had a stressor within a year, most commonly a move to a new home or school. In another, 91% had at least one adverse stressor, most often bereavement or illness of a close person (35%) or separation from a significant figure (32%).

Diagnostic yield

In the 164-patient cohort, DSRD patients more often had positive ANA, low complement C3, abnormal cytokines, and elevated ferritin than DS-only controls. EEG was abnormal in 30%, MRI in 33%, and CSF in 21%. Thyroperoxidase antibody titers were elevated in 91% of regression cases versus 23% of controls in the Worley series.

The immune substrate

Trisomy 21 triples the dose of the interferon receptor gene cluster, producing persistent elevation of up to 22 cytokines at steady state, expansion of CD11c+ B cells, and hundreds of autoantibodies including CNS-directed species. Murine correction of interferon receptor copy number normalized antiviral responses and improved cognition. Within DSRD specifically, trio exome sequencing of 41 patients found de novo variants in immune regulatory genes in 20%, with four pathogenic or likely pathogenic variants all implicated in type I interferon signaling; carriers were more likely to have preceding triggers, decline within one month, and MRI abnormalities.

A tempting wrong turn

Alzheimer's disease is almost never the answer before 40

DS-associated Alzheimer's disease is now understood as a genetically determined form of AD analogous to autosomal dominant AD, with near-universal amyloid and tau neuropathology by age 40. This creates a persistent temptation to invoke it. The temptation should be resisted in the 10 to 30 age band.

  • 1

    Clinical dementia prevalence in DS adults aged 30 to 39 was 0% using validated diagnostic instruments, rising to 8.9% at 45 to 49 and 32.1% at 55 to 59.

  • 2

    In the 388-patient multimodal biomarker cohort, median age at prodromal AD diagnosis was 50.2 years and at AD dementia 53.7 years. Meta-analytic means are 52.7 and 56.2 years.

  • 3

    The biomarker cascade begins in the third decade: CSF A-beta 42/40 diverging at about age 28, plasma NfL at about age 30, amyloid PET in the late 30s, plasma p-tau217 at about 37.3 years.

  • 4

    Because pathology tracks age so tightly in DS, age alone was as sensitive as plasma p-tau217 for detecting preclinical pathology, which is why a negative p-tau217 in a 20-year-old adds little that chronological age had not already supplied.

  • 5

    When decline does emerge in the 30s, it more often manifests as behavioral and personality change, apathy, impulsivity, and executive dysfunction, than as classic amnesia. This overlaps heavily with DSRD.

  • 6

    A first unprovoked seizure after age 40 to 45 is highly suggestive of symptomatic AD, with up to 75% of DS adults with dementia developing epilepsy. That feature is absent in the DSRD age range and therefore useful when present.

Highest yield

Catatonia and delirium

Catatonia

Catatonia is the highest-yield finding in this differential because it is common, frequently missed, and treatable. In the 164-patient cohort it was significantly more likely in DSRD than in alternative diagnoses. Across nine DSRD patients with catatonia, mean BFCRS was 17.3 with 11.1 positive signs; staring was universal, and mutism, grimacing, and rigidity each occurred in about 78%.

The APA resource document names DSRD as an exemplar of catatonia presenting as deviation from neurodevelopmental baseline, and stipulates that only features differing from the patient's ordinary baseline should be counted, otherwise chronic intellectual disability features generate false positives. A 14-item BFCRS screen with a threshold of 2 or more triggers the full 23-item scale.

Lorazepam challenge: 2 mg IV (1 mg in children or those at risk of respiratory compromise), with a 50% or greater BFCRS reduction defining a positive response, reassessed at 30 minutes for parenteral and 2 hours for oral dosing. In a 54-patient pediatric multisite series, mean BFCRS fell from 16.6 to 9.5 with a large effect size. Sedation alone does not exclude catatonia; patients may show a positive response on awakening.

Delirium

Delirium must be co-assessed rather than excluded by fiat. Among 136 critically ill patients, 31% had both catatonia and delirium, 43% delirium alone, and 3% catatonia alone; among 205 medical inpatients with delirium, 30% had concurrent catatonia. DSM-5-TR formally precludes diagnosing catatonia in the presence of delirium, but ICD-11 and the APA document recognize co-occurrence and recommend simultaneous management.

Delirium features fluctuating and generally reduced alertness, whereas catatonia features mutism and stupor with cognition relatively preserved though inaccessible. Lorazepam response also discriminates: 60% of patients with isolated catatonia responded, versus 1 of 9 with comorbid catatonia and delirium.

Malignant catatonia, catatonia with dysautonomia and hyperthermia, often with elevated CK, leukocytosis, and low serum iron, is a medical emergency, and many authorities regard neuroleptic malignant syndrome as its drug-induced variant, sometimes clinically indistinguishable.

Reversible causes first

Medical mimics that must be excluded

ConditionBase rate in Down syndromeClinical clueTest
Thyroid disease / Hashimoto encephalopathyThyroid disorder is common in Down syndrome; SREAT is reported and may be euthyroid.Rapidly progressive cognitive decline, seizures, hallucinations, behavioral change; often steroid responsive.TSH, free T4, anti-TPO and anti-thyroglobulin even if TFTs are normal; EEG.
Obstructive sleep apnea50 to 96% of children, over 90% of adults.AHI correlates with behavior problems and lower developmental quotient; adults present with sleepiness, depression, and mood dysregulation rather than sleep complaints.Polysomnography, among the most frequently abnormal studies in regression cohorts.
Celiac disease5 to 11% prevalence in Down syndrome.Behavioral change and GI symptoms; 9% of regression referrals with an alternative diagnosis.TTG-IgA with simultaneous quantitative IgA (IgA deficiency is common).
Vitamin D and B12 deficiency77% vitamin D deficient in one pediatric cohort, 32% severely.B12 deficiency can cause psychosis, hallucinations, personality change, and delirium in any population, and is reversible when treated early.25-OH vitamin D, serum B12.
Atlantoaxial instability / cervical myelopathyRadiographic instability 8 to 13%; symptomatic cord compression 1 to 2%.Gait change, hand dysfunction, incontinence, and reduced activity misread as regression.Neurologic exam each visit for gait, tone, reflexes, clonus; imaging if symptomatic.
Moyamoya syndrome26-fold increased risk versus the general population.Stroke-predominant presentation with longer diagnostic delay; rising blood pressure percentiles may precede by 6 to 24 months.MRI/MRA with a low threshold, DSA if positive.
Epilepsy / nonconvulsive statusRoughly 8% seizures, up to 13% lifetime epilepsy; 40% with onset in the third decade or later.Post-ictal dysphoria and irritability; NCSE is hard to detect against an abnormal baseline EEG.EEG (abnormal in 30% of regression cases).
Constipation and occult painFunctional constipation 34 to 36%; 52% meet Rome IV criteria for a gut-brain disorder.Pain expressed as agitation, self-injury, or apparent regression in limited-communication patients.Abdominal exam, bowel history, medication review.
Hearing and vision impairmentHearing loss up to 75 to 85%; eye problems 60 to 80%.Undetected sensory loss mimics withdrawal and cognitive decline.Audiometry, ophthalmologic evaluation.
Medication effectPolypharmacy is common in intellectual disability crisis admissions.Patients respond to lower doses and are more side-effect prone; antiepileptics, stimulants, and anticholinergics carry high behavioral risk.Full medication reconciliation.

The 2022 consensus workup reached 100% agreement on bloodwork and urine studies, 96% on neuroimaging and other studies (EEG, polysomnography), and 88% on lumbar puncture. The most frequently abnormal findings in the international database were 25-OH vitamin D, polysomnography, thyroid peroxidase antibodies, and celiac screens.

Therapy as a test

Treatment response as a diagnostic instrument

Because no biomarker confirms DSRD, differential response to therapy carries real diagnostic weight.

Immunotherapy response

Argues for DSRD with an inflammatory substrate

In the largest prospective IVIg study (n = 82), scores improved across all three measures (BFCRS mean difference -6.68, CGI-S -1.27, NPITS -6.50). After a standardized wean at 9 to 12 months, 46% relapsed, predicted by any abnormal neurodiagnostic study, abnormal MRI, abnormal LP, and a personal autoimmunity history. With both MRI and LP abnormal, the odds of immunotherapy responsiveness approached 100%.

Benzodiazepine response

Argues for catatonia

Lorazepam response in catatonia generally runs 66 to 100%. Dosing creates real tension: Down syndrome patients are often side-effect sensitive and respond to lower doses, yet catatonia in intellectual disability and autism frequently requires 6 to 24 mg/day. Maximal response to scheduled lorazepam typically appears within 3 to 7 days.

ECT response

Argues for catatonia or affective illness

ECT most dramatically restored baseline in the Miles series, though recurrence often required additional courses. A DSRD case with only a small unsustained IVIg response converted to a rapid sustained response once ECT was added. Across intellectual disability broadly, 84% showed effectiveness without important adverse effects.

Antipsychotic failure or intolerance

Argues against primary psychosis

In the French DSRD series, SSRIs, atypical antipsychotics, and benzodiazepines were ineffective and poorly tolerated. This is a safety issue as much as an efficacy one: NMS was three times more common in adults with intellectual disability prescribed antipsychotics, and neurodevelopmental disorder was the strongest independent NMS risk factor in a national cohort of over one million antipsychotic users aged 5 to 24 (HR 7.11).

Putting it together

The diagnostic sequence

  1. 1

    Confirm the phenomena are not developmentally normative self-talk or imaginary companionship, by establishing change from baseline plus functional cost.

  2. 2

    Screen for catatonia with the BFCRS using baseline-referenced scoring, and perform a lorazepam challenge, which is simultaneously diagnostic and therapeutic.

  3. 3

    Assess concurrently for delirium, recognizing roughly 30% co-occurrence.

  4. 4

    Complete the consensus tiered workup: thyroid with antibodies, celiac with quantitative IgA, vitamin D and B12, ANA / complement / ferritin / cytokines, polysomnography, EEG, MRI with or without MRA, and lumbar puncture including paired serum and CSF NMDAR antibodies.

  5. 5

    Treat DSRD as the leading hypothesis if the tempo is subacute, a psychosocial trigger is identifiable, and catatonia is present, recognizing that about 20% of such referrals prove to be something else.

  6. 6

    Do not invoke Alzheimer's disease, for which clinical prevalence is 0% at ages 30 to 39 in validated series.

  7. 7

    Reserve antipsychotics for genuinely established primary psychosis, at low doses, given the elevated NMS risk.

Honest limits

Evidence gaps

  • No randomized controlled trial of any intervention for DSRD, catatonia in Down syndrome, or psychosis in Down syndrome has been published.
  • Incidence remains unknown; a single Italian cohort estimated 7.1% among 70 adults.
  • It is unresolved whether DSRD is a single entity or a final common pathway for immune, psychiatric, and stress-related processes.
  • IVIg dosing, duration, and weaning protocols remain empirical.
  • The BFCRS has not been formally validated in intellectual disability despite universal use in this literature.
  • No delirium instrument is validated for non-ICU adults with intellectual disability.
  • The mechanism of Down syndrome's apparent relative protection against schizophrenia is unexplained.
  • The empirical boundary between benign self-talk and prodromal psychosis in Down syndrome lacks any validated criterion set.

Synthesized from the peer-reviewed literature on Down syndrome regression disorder, catatonia, and psychosis in intellectual disability, including Rosso et al. (2020), Santoro et al. (2020 to 2026), Worley et al. (2015), Bonne et al. (2023), Bull (2020), Tsou et al. (2020), Fortea et al. (2020, 2021), and Wilson et al. (2025). This page is clinical education, not medical advice.

Keep reading

All clinician tools pages →

This page was medically reviewed by Eric Wexler M.D., Ph.D. on August 14, 2026.