The research team, led by Tao Chen, Qingyu Sun, and senior author Muireann Irish, analysed data from 432 people diagnosed with clinically probable frontotemporal lobar degeneration (FTLD) or Alzheimer’s disease between 2008 and 2025 at the Brain and Mind Centre in Sydney. Of these, 358 individuals had complete carer burden data and formed the primary analytic cohort, comprising 230 people in the FTLD group and 128 in the Alzheimer’s disease group. The FTLD group spanned the full clinical spectrum, including behavioural variant frontotemporal dementia, left- and right-sided semantic dementia, progressive nonfluent aphasia, progressive supranuclear palsy, and corticobasal syndrome, while the Alzheimer’s group included typical presentations and logopenic progressive aphasia. Diagnoses were reached by multidisciplinary consensus using internationally recognised criteria, and all participants scored at least 40 out of 100 on a standard cognitive screening battery at baseline.

The methodological design of the study is what sets it apart from earlier work. Rather than relying on the Neuropsychiatric Inventory, the instrument used in most previous caregiver studies, the team turned to the Cambridge Behavioural Inventory—Revised, a carer-reported questionnaire whose broader coverage captures behavioural domains that other tools routinely miss. Chief among these is behavioural rigidity, an umbrella term encompassing stereotypies such as excessive hand rubbing, compulsions like relentless hand washing, stereotyped catchphrases, impulsive acts including pathological gambling, hoarding, restricted interests, insistence on sameness, and ritualistic routines. Carer burden itself was measured with the widely used 12-item Zarit Burden Interview, which quantifies the emotional, physical, and role-related demands of caregiving on a scale from 0 to 48.

The statistical approach was equally rigorous. Multivariate regression models predicted carer burden from seven behavioural and psychological symptom domains, adjusting for patient age, sex, education, disease duration, disease severity, and diagnostic category. Missing data were handled through multiple imputation by chained equations, and the team complemented standardised regression coefficients with relative importance metrics based on the Lindeman–Merenda–Gold method, which partitions explained variance among correlated predictors. This dual strategy allowed the researchers to distinguish symptoms that independently drive burden from those that contribute through shared, synergistic effects with co-occurring behaviours, a nuance that single-coefficient analyses cannot capture.

The headline finding is striking in its clarity. Across the entire dementia cohort, two symptoms emerged as significant predictors of carer burden: behavioural rigidity, with a standardised coefficient of 0.15, and apathy, indexed by the motivation subscale, with a coefficient of 0.14. Both were modest in absolute terms but statistically robust, with confidence intervals excluding zero. Crucially, global cognition, measured by the Addenbrooke’s Cognitive Examination, showed no significant association with carer burden whatsoever. This dissociation is particularly provocative in Alzheimer’s disease, where cognitive decline is conventionally assumed to be the primary engine of carer distress. The data suggest instead that it is the disruption of motivation, behaviour, and daily routine, not the erosion of memory, that most corrodes carer wellbeing.

When the analyses were run separately within diagnostic groups, a syndrome-specific pattern emerged. In the FTLD group, behavioural rigidity was the strongest and most important predictor of carer burden, with a standardised coefficient of 0.18 and the largest relative importance score of 0.137. This finding held even when the researchers excluded participants with behavioural variant frontotemporal dementia, the syndrome most obviously associated with rigid behaviour, indicating that the effect extends across the broader FTLD spectrum. The authors propose a compelling mechanistic account: rigid behaviours are typically high-frequency and resistant to change, forcing carers to organise entire days around accommodating rituals and rules. Over time, this does not merely add to the caregiving workload; it fundamentally reconfigures the carer’s own life, constraining the timing, sequencing, and nature of everyday activities at the expense of efficiency and shared decision-making.

In the Alzheimer’s group, the picture shifted. Here, apathy and abnormal behaviour were the significant predictors, each with a standardised coefficient of 0.26, and apathy carried the largest relative importance value of 0.232. The findings align with qualitative research suggesting that apathy imposes its toll through a subtle but relentless reconfiguration of the caregiving relationship. When a person with dementia loses the drive to initiate activity, responsibility for starting and sustaining every part of daily life falls to the carer, who must navigate the tension between gently steering engagement and preserving the person’s remaining autonomy. That balancing act, the literature suggests, breeds frustration, guilt, and a gradual, painful acceptance of diminished reciprocity in the relationship.

Perhaps the most forward-looking element of the study is its longitudinal component. In 172 participants assessed at baseline and again roughly one year later, the researchers examined how within-person changes in symptoms related to changes in carer burden over the same interval. A single domain stood out: mood-related symptoms, including irritability and agitation, were the only significant longitudinal predictor of increasing carer burden across the combined cohort, with a standardised coefficient of 0.24. No significant longitudinal predictors emerged within either diagnostic group alone, likely reflecting the modest sample sizes and heterogeneous progression rates of individual syndromes. The transdiagnostic pattern, however, is clinically meaningful. Mood-related symptoms fluctuate with situational demands and are notoriously difficult to anticipate, and experience-sampling research has shown that unpredictable and uncontrollable events are potent drivers of negative affect in dementia carers. Even infrequent episodes of agitation, previous work has demonstrated, impose disproportionately high levels of stress, because each episode forces carers to abandon coping strategies that had been working and reallocate their emotional resources.

The clinical implications are direct and actionable. The authors argue that routine screening for behavioural rigidity and apathy should be considered in all patients with dementia, with psychoeducation at the point of diagnosis preparing carers, particularly in Alzheimer’s disease, for apathy as a prominent and burdensome feature rather than a secondary annoyance. In the absence of effective disease-modifying therapies, symptom-focused management may offer the greatest realistic potential for reducing carer burden. Given the well-documented dysfunction of serotonergic and dopaminergic systems in FTLD, and their established links to cognitive flexibility, pharmacological modulation of these pathways may eventually help alleviate rigidity and repetitive behaviours, though current guidelines recommend prioritising non-pharmacological approaches whenever symptoms are mild and pose minimal risk. The study is not without limitations: carer-side factors such as personality, coping style, social support, and hours of care were not modelled, pathological confirmation of diagnosis was unavailable, and the functional rating scale used to index severity has not been validated across every syndrome included. Larger, multi-centre samples with extended follow-up will be needed to refine these syndrome-specific relationships. Yet the core message stands firm. What determines whether a carer buckles under the strain is not how far memory has faded, but whether rigid routines have colonised the household, whether motivation has drained away, and whether mood disturbances are flaring without warning. Recognising these targets, and treating them as the primary clinical outcomes they truly are, may be the most effective way to protect the people who hold dementia care together.

Subject of Research: Behavioural and psychological symptom determinants of carer burden in frontotemporal lobar degeneration and Alzheimer's disease

Article Title: Behavioural and psychological symptom determinants of carer burden across clinical syndromes associated with frontotemporal lobar degeneration and Alzheimer’s disease: a transdiagnostic analysis

Article References: Behavioural and psychological symptom determinants of carer burden across clinical syndromes associated with frontotemporal lobar degeneration and Alzheimer’s disease: a transdiagnostic analysis. (n.d.). https://doi.org/10.1007/s00415-026-14132-1

Image Credits: AI Generated

DOI: 10.1007/s00415-026-14132-1

Keywords: carer burden, dementia, frontotemporal lobar degeneration, Alzheimer's disease, behavioural rigidity, apathy, behavioural and psychological symptoms, Zarit Burden Interview, Cambridge Behavioural Inventory-Revised, caregiving, longitudinal analysis, neuropsychiatry

The review, led by Jim Jacob and Federica Sarno of the University of Groningen’s University Medical Center Groningen, together with Maria Zernova, Inge M. Strating, Barbara van Munster, and Monika Trzpis, set out to answer a deceptively simple question. If delirium arises from an interaction between predisposing vulnerabilities—advanced age, neurodegeneration, frailty—and acute precipitating factors such as inflammation, surgery, or critical illness, could methylation changes be part of the molecular wiring that connects the two? The team systematically screened the literature and identified 1,110 potentially relevant articles. After applying their inclusion criteria, only nine studies survived. That attrition alone tells a story: despite a strong theoretical rationale, the empirical record linking DNA methylation to delirium is thin, and the authors conclude that robust evidence remains lacking.

The nine included studies split into two broad families. Six examined differential DNA methylation at candidate CpG sites—the cytosine–guanine dinucleotides where methyl groups are most commonly attached and where methylation is known to influence transcription. Three took a different tack, assessing age-related methylation patterns stratified by delirium status, exploiting the fact that the methylation state of many sites across the genome shifts predictably with chronological and biological age. Only four of the nine studies performed true genome-wide differential methylation analysis, the kind of unbiased survey that can, in principle, discover methylation signatures no one thought to look for in advance. The remaining studies were hypothesis-driven, focusing on genes selected for their plausibility.

Among the four genome-wide studies, two independent investigations each identified a single CpG site that was hypomethylated in patients who experienced delirium. The first site, cg21295729, annotated to the gene LDLRAD4, was found in blood and involved a comparison of 43 delirium cases against 44 controls. The second, cg16526133, annotated to ADAMTS9, emerged in brain tissue, comparing 11 cases with 25 controls. Hypomethylation at a CpG site typically suggests increased accessibility of DNA to transcriptional machinery, hinting—though only hinting—that these genes might behave differently in people who develop the syndrome. Notably, LDLRAD4 and ADAMTS9 have no obvious shared functional storyline, and the two studies pointed to different tissues, different cohorts, and different design logic. The other two genome-wide studies found no statistically significant methylation differences at all, a result the review does not shy away from reporting.

Perhaps the most intriguing signal in the review involves inflammation. Gene-specific analyses showed that methylation levels at CpG sites annotated to the TNF gene—the canonical inflammatory cytokine tumor necrosis factor—correlated negatively with age in peripheral blood mononuclear cells and in whole blood samples taken from delirious patients. In plain terms, the older the delirious patient, the lower the methylation at these TNF-associated sites, a pattern consistent with progressively disinhibited inflammatory gene expression in a vulnerable host. Critically, this age-dependent pattern was absent in saliva and buccal samples, and absent in control subjects. That tissue specificity matters enormously. It suggests that the association is not a generic artifact of aging, but something particular to blood-derived immune cells in patients who actually went through a delirious episode. Blood, of course, carries the immune cells believed to orchestrate the systemic inflammatory response that many theories place at the center of delirium pathophysiology.

A complementary pattern emerged for genes involved in neurotrophic signaling. CpG sites annotated to BDNF, GDNF, and NR4A2—genes encoding factors that support neuronal survival, plasticity, and synaptic resilience—showed a positive correlation with age in delirious patients, meaning methylation increased as patients got older. Because increased promoter methylation is often, though not invariably, associated with reduced gene expression, the finding raises the possibility of an age-related epigenetic tightening of the brain’s own maintenance systems. In a patient facing an acute inflammatory or metabolic insult, the combination of a revved-up immune program and a dampened neurotrophic program is precisely the kind of double hit that theories of delirium have long hypothesized: the brain, already running on diminished reserves, is simultaneously provoked and deprived of support at the moment of the insult.

The reviewers are careful, and admirably so, about what such signals can and cannot support. Methylation measured in blood may or may not reflect what is happening in neurons or glia; the brain-tissue study was small, with just 11 cases; and methylation is highly sensitive to cell-type composition, meaning that shifts in the proportions of immune cell subsets could masquerade as differential methylation. The studies also varied widely in sample size, tissue type, statistical thresholds, and whether analyses corrected for the many confounders—age, medication, illness severity, smoking, diet—that are known to sculpt the methylome. Two of the four genome-wide studies reported nothing significant, and the authors emphasize that the existing literature is limited both in number and in methodological rigor, constraining firm conclusions about whether methylation is a cause, a consequence, or merely a correlate of the syndrome.

What makes the field worth watching is the underlying biology. DNA methylation is a dynamic layer of regulation; while many marks are stable, others respond within hours or days to inflammation, hypoxia, and stress hormones, all of which surge during the kinds of acute events that precipitate delirium. Methylation also offers something epidemiology alone cannot: a mechanistic bridge between vulnerability and insult. If older brains carry methylation signatures that favor TNF derepression in immune cells while simultaneously suppressing neurotrophic support, then a surgical trigger or a fever could plausibly flip a network of gene-expression switches that cumulatively destabilize cognitive control circuits. The convergence of the two methylation trends—one pointing toward immune activation, the other toward weakened neuronal maintenance—is conceptually coherent with prevailing models of delirium, even if the empirical evidence has not yet reached the strength that coherence deserves.

The practical implications of the review are correspondingly modest but clear. The authors argue that what the field needs is larger, well-powered, longitudinal cohort studies with genome-wide methylation profiling, careful adjustment for confounders, and ideally paired blood and brain tissue to test whether peripheral signals mirror central ones. Prospective designs, in which methylation is measured before surgery or before an infection takes hold, would help distinguish pre-existing epigenetic vulnerability from methylation changes induced by the episode itself. Standardized covariate adjustment and cell-composition correction would strengthen comparability across studies. Until then, the review stands as both a map and a warning: DNA methylation remains a theoretically compelling contributor to delirium pathophysiology, supported by a handful of suggestive and partially converging findings, but the field has yet to produce the rigorous, replicated evidence required to elevate epigenetics from plausible mechanism to established biology. For clinicians and researchers watching a rapidly aging global population, closing that evidence gap is not an academic luxury; it is the necessary groundwork for the first epigenetically informed approaches to predicting, and perhaps one day preventing, one of medicine’s most disruptive syndromes.

To appreciate why methylation has drawn attention in delirium research, it helps to understand how plastic this mark can be. Cytosine methylation is maintained by enzymes of the DNMT family and actively removed through TET-mediated oxidation pathways, creating a reversible system that responds to environmental and physiological cues. In the immune system, stimulus-responsive methylation changes are well documented: monocytes and T cells remodeled their methylomes within hours of activation, and inflammatory cytokine promoters are among the genomic regions most sensitive to such remodeling. This makes blood an attractive, if imperfect, surrogate tissue for studying the interface between systemic inflammation and brain dysfunction.

Age adds a second layer of complexity. Large consortium analyses of thousands of blood and brain samples have shown that methylation drifts in a highly stereotyped way across the lifespan, and that some loci accelerate or decelerate their drift in the context of disease. The delirium studies reviewed here sit squarely within this tradition, testing whether the syndrome modifies the expected age-methylation relationship rather than producing absolute differences between cases and controls. That analytic framing may be well suited to a condition defined by interactions between age-related vulnerability and acute insult, though it also demands large, age-diverse samples that most existing cohorts cannot provide.

The review also reflects a broader methodological moment. Epigenome-wide association studies in neuropsychiatry have repeatedly shown that small cohorts underpowered to detect the modest effect sizes typical of methylation differences. The field’s experience elsewhere suggests that combining cohorts, harmonizing platforms, and sharing raw data will be prerequisites for any credible replication of the signals highlighted in this work.

Subject of Research: DNA methylation as a candidate molecular mechanism in delirium pathophysiology

Article Title: A systematic review on the contribution of DNA methylation to delirium pathophysiology

Article References: Jacob, J., Zernova, M., Strating, I. M., van Munster, B., Sarno, F., & Trzpis, M. (2026). A systematic review on the contribution of DNA methylation to delirium pathophysiology. Epigenetics Communications. https://doi.org/10.1186/s43682-026-00051-9

Image Credits: AI Generated

DOI: 10.1186/s43682-026-00051-9

Keywords: delirium, DNA methylation, epigenetics, systematic review, TNF, inflammation, aging, neurotrophic signaling, BDNF, genome-wide methylation, neuropsychiatry, blood-based biomarkers