A new sweet tooth, sudden binge eating, or a compulsion to put non-food objects in the mouth cannot diagnose frontotemporal dementia, because changed eating is only one item on a checklist that requires several. Frontotemporal dementia (FTD) is a group of disorders caused by damage to the frontal and anterior temporal lobes, and its behavioural variant (bvFTD) is defined by a pattern of changes, not by any single one. The eating symptom matters, and it is worth reporting to a doctor. But on its own it appears in fewer than half of people who have not yet developed other features, appears in conditions that are not FTD at all, and cannot reach even the lowest diagnostic tier of the criteria used worldwide.
Medical information disclaimer: This article is for general educational purposes only and does not provide medical advice, diagnosis, or treatment. Always consult a physician or other qualified health professional about symptoms, medications, tests, or treatment decisions.
Table of Contents
- Three of six features, before "possible" is even on the table
- How often does eating change actually occur in FTD?
- Everything else that changes how someone eats
- Why this goes wrong so often, and for whom
- What a clinician actually does, and what to bring
- Frequently Asked Questions
Three of six features, before "possible" is even on the table
The 2011 international consensus criteria, published by Rascovsky and colleagues in *Brain*, list six early behavioural or cognitive features of bvFTD: disinhibition, apathy or inertia, loss of empathy, perseverative or compulsive behaviour, hyperorality and dietary change, and a dysexecutive neuropsychological profile — poor planning, judgement, and mental flexibility with relatively preserved memory. The criteria require three of those six before a clinician can record even "possible bvFTD." "Hyperorality" means an altered relationship with putting things in the mouth: craving sweets or carbohydrates, eating far past fullness, drinking or smoking more, or mouthing inedible objects. It counts once.
It does not count three times because it is dramatic. "Possible" is also the weakest of the tiers. Rascovsky's group set out in *Brain* that "probable bvFTD" additionally requires documented functional decline plus frontal or anterior temporal atrophy or hypometabolism on a scan — shrinkage or reduced activity in those regions. An eating change with no imaging support cannot arrive at that tier by any route.
How often does eating change actually occur in FTD?
Less often than its reputation suggests, at the point where it would be most useful. Morrow and colleagues, writing in the *Journal of Neuropsychiatry and Clinical Neurosciences*, report that hyperorality is present in roughly 50% of early-stage bvFTD patients, rising to about 63–75% in advanced disease. That cuts both ways. Half of people with early bvFTD do not have it, so normal eating is not reassurance — families sometimes wait for a "classic" sign that never arrives.
And because it is not universal even among confirmed cases, its presence cannot carry the diagnosis by itself. The full checklist is not perfect either. In the validation cohort of pathology-confirmed cases, the revised criteria reached about 86% sensitivity for possible bvFTD and 76% for probable. If the complete multi-feature assessment misses a meaningful minority of true cases, one symptom inspected alone is a far weaker instrument.
Everything else that changes how someone eats
Altered appetite is common in later life and mostly means something other than FTD. Alzheimer's disease more typically brings decreased appetite, early swallowing difficulty, and weight loss, though increased appetite also occurs — so the direction of change does not cleanly separate the two diseases.
Research summarised in the *Journal of dementia and Alzheimer's Disease* found appetite disorders were more common in older adults with frontotemporal atrophy who did not have dementia at all. Ordinary explanations deserve to be ruled out first, and often are the answer: There is also a "phenocopy" syndrome. A 2019 review by Valente and colleagues in *Alzheimer's Research & Therapy* describes people with bvFTD-like behaviour for years alongside normal MRI and FDG-PET scans and no progression to dementia, estimated at 7–37% of bvFTD-like presentations, with psychological and psychiatric factors implicated.
- New medication, particularly antidepressants, antipsychotics, steroids, or diabetes drugs
- Depression, anxiety, or bereavement, which can raise or flatten appetite
- Thyroid disease, poorly controlled diabetes, or vitamin deficiency
- Dental pain, ill-fitting dentures, or loss of taste and smell
- Alcohol use, sleep disorders, or untreated sleep apnoea
Why this goes wrong so often, and for whom
The people most affected are not the ones most doctors are watching for dementia. The National Institute on Aging reports that about 60% of people with FTD are aged 45 to 64 — the working-age band where a new sweet tooth, late-night bingeing, or reckless spending is read as stress, burnout, or a midlife mood disorder. The numbers on misdiagnosis are stark. In a 2025 caregiver survey published in *Alzheimer's & Dementia*, Milliard and colleagues found 52% of patients were initially misdiagnosed — depression in 73% of those, other psychiatric diagnoses in 58%, anxiety in 50%, mild cognitive impairment in 38%.
Three-quarters waited more than a year for a diagnosis, and two-thirds saw three or more doctors. The error runs in both directions, which is the argument for a full assessment rather than a shortcut. Treating a single eating change as proof labels people who have a treatable depression or a phenocopy syndrome. Dismissing it as stress delays the people who do have FTD by years.
What a clinician actually does, and what to bring
The NIA states plainly that there is no single diagnostic test for FTD. Diagnosis is built from a symptom history plus the exclusion of other conditions, using blood tests, sleep studies, neuropsychological testing, and brain scans. The most useful thing a family can supply is trajectory, not a list of odd moments.
Writing in *Alzheimer's & Dementia: Diagnosis, Assessment & Disease Monitoring*, Dijk and colleagues note that bvFTD symptoms progress while primary psychiatric symptoms are often stable, and that social cognition — especially recognising emotion in other faces — is more impaired in bvFTD. Structured instruments such as DAPHNE and the FTD Module help clinicians work through that differential. Practical preparation for an appointment:.
- Date the change. When did eating shift, and has it worsened, held steady, or come and go?
- Write down the other five features separately: tactlessness or disinhibition, withdrawal or apathy, reduced warmth toward people they love, repetitive rituals or phrases, and poor planning or judgement.
- Bring a complete medication list, including anything started in the six months before the change.
- Ask for a referral to neurology or a specialist cognitive clinic if the GP's workup is clear and the behaviour is progressing.
- If depression is diagnosed, agree on a review date. Psychiatric symptoms that do not stabilise with treatment, and keep broadening, are the pattern worth re-examining.
Frequently Asked Questions
My relative suddenly craves sugar constantly. Should I ask about FTD?
Raise it, but describe it as one change among others rather than as a suspected diagnosis. Ask the doctor to rule out medication effects, thyroid problems, diabetes, and depression, and mention any change in tact, empathy, planning, or repetitive behaviour.
Does normal eating mean it is not FTD?
No. Morrow and colleagues report hyperorality in only about half of early-stage bvFTD patients, so appetite can stay entirely unchanged in someone who has the disease.
Can a brain scan settle it on its own?
Not alone. Imaging is required for "probable bvFTD" but only alongside three of the six clinical features and documented functional decline, and phenocopy cases show bvFTD-like behaviour with normal scans.
What if the eating change turns out not to be progressing?
That is meaningful information. Stability over time points away from bvFTD and toward a psychiatric or phenocopy explanation, which is why clinicians ask for a dated history rather than a snapshot.
