Walk around almost any city and men in their forties or fifties with large bellies are easy to spot. Move into the seventies and eighties, however, and men who look dramatically large seem much rarer.
That observation is not entirely imaginary.
But the one-line explanation — “they all died before getting old” — is much too crude.
Japan’s latest national data show something more interesting.
1. The short answer: “overweight” remains common, but the very high-BMI tail gets thinner
In Japan’s 2024 National Health and Nutrition Survey, the share of men with BMI at or above 25 was 35.8% in their forties, 38.3% in their fifties, 33.6% in their sixties, and 29.4% among men aged 70 or older.[1][2]
So nearly three in ten men over 70 were still in the BMI 25-plus category.
Older men with higher BMI have not vanished.
The picture changes when we look at BMI 30 or above:
- 40–49: 7.7%
- 50–59: 7.2%
- 60–69: 3.5%
- 70+: 2.4%
- 70–74: 3.0%
- 75–79: 2.5%
- 80–84: 2.3%
- 85+: 0.7%
BMI 30-plus is about one in fourteen men in their fifties, one in forty-two among men over 70, and roughly one in 140 among men 85 or older.
You cannot estimate BMI by looking at strangers. Still, the part of the distribution most likely to look conspicuously large really does become much smaller.
The mildly heavy background characters remain. The screen-filling boss characters become rare.
2. Average body weight is also lower in older age groups
In the same 2024 survey, average male body weight was 70.9 kg in the fifties, 68.1 kg in the sixties, 63.3 kg at age 70-plus, and 60.5 kg at age 80-plus. Mean BMI was 24.4, 23.9, 23.4 and, among men 85-plus, 22.7.[2]
But there is a crucial statistical warning.
These are cross-sectional age groups.
The survey did not take one man at age 50 and follow him until 85.
A cross-sectional table is a group photo, not security-camera footage of the same people over thirty years.
Age effects and birth-cohort effects are mixed together.
3. Why do very large older men become less visible?
There is no single cause. At least five mechanisms can overlap.
3-1. Survival selection is real, but it is not the whole story
Higher BMI is associated with several diseases, and pooled Japanese cohort data have found higher all-cause mortality in the BMI 30–40 range compared with a BMI 23–25 reference group.[3]
So survival selection can contribute to who reaches advanced age.
But that does not mean most heavy people die before 70.
A Japanese cohort of adults aged 65–79 found a clear mortality disadvantage at low BMI, while the broad BMI range of 20–29.9 had the lowest all-cause mortality.[4]
Old age is not a game in which “lower BMI always wins.”
3-2. Some people lose weight as they age
Illness, reduced appetite, swallowing or dental problems, medications, social isolation and changes in activity can all contribute to weight loss in later life.
A meta-analysis in older adults found weight loss to be associated with higher all-cause mortality compared with stable weight.[5]
So when an 80-year-old becomes thin, that is not automatically a health achievement.
Unintentional weight loss can be a warning light rather than a victory screen.
3-3. Muscle loss changes the silhouette
Muscle mass and strength commonly decline with age.
When excess fat coexists with reduced muscle mass or function, the condition is discussed as sarcopenic obesity. ESPEN and EASO have proposed a consensus definition and diagnostic procedure.[6]
This creates a visual trap.
A person may retain substantial body fat while losing muscle from the shoulders, chest and legs. The body may therefore look less massive even when body composition is still metabolically unfavorable.
The number on the scale is not the whole character sheet.
3-4. Birth cohorts are different
An 85-year-old today and a 50-year-old today did not grow up in the same food environment.
Analysis of Japanese national surveys from 1973 to 2016 found major long-term increases in overweight and obesity among older men, with differences by birth cohort.[7]
That means today’s fifty-somethings cannot simply be pasted into the body-size distribution of today’s eighty-somethings.
Future older populations may look different.
3-5. “Older people seen on the street” are not all older people
This is a sampling problem.
People seen at stations, supermarkets, restaurants and parks are disproportionately those able to go out independently.
People who are hospitalized, living in institutions, largely homebound or limited in mobility are less visible in a casual street sample.
The human eye is not a national census.
One hundred people seen near a station do not constitute a representative epidemiological survey.
4. “I do not see them, therefore they all died” compresses five processes into one
Mortality selection matters.
But so do:
survival selection, weight loss, loss of muscle, birth-cohort differences, and selection into public visibility.
In game terms, not every missing character was defeated.
Some left the map. Some changed class and appearance. Some belong to a different generation of the game. Some simply do not spawn in the area you are walking through.
Reducing all of that to one “death” button is bad log analysis.
5. In older adults, “heavier” and “more dangerous” are not perfect synonyms
Low BMI and weight loss can themselves matter in later life.
A 2025 Japanese study of more than 315,000 adults aged 65–74 found that low BMI and a higher A Body Shape Index were independently associated with higher all-cause mortality. The authors concluded that combining body-shape information with BMI was more informative than relying on BMI alone.[8]
For older adults, it is therefore useful to separate:
how much fat is present, how much muscle is preserved, where fat is distributed, whether weight has recently fallen unintentionally, whether the person can eat adequately, and whether the person can move and function.
BMI is a useful map. It is not an aerial photograph.
6. The street observation is surprisingly sharp — the causal story is not
The impression that extremely large men are uncommon after 70 is directionally consistent with Japan’s 2024 data.
The BMI 30-plus share falls from 7.2% among men in their fifties to 2.4% at 70-plus and 0.7% at 85-plus.[2]
But the inference chain
“I rarely see them” → “they do not exist” → “they all died”
jumps three steps at once.
The observation can be right while the explanation is incomplete.
7. The funniest statistical twist: plenty of 70-plus men are still above BMI 25
The eye tends to remember extremes.
Yet 29.4% of Japanese men aged 70-plus were still in the BMI 25-plus category in the 2024 survey.[1][2]
What shrinks most dramatically is not the entire higher-BMI population, but the conspicuously large BMI 30-plus tail.
That is why both statements can feel true:
“I hardly ever see very fat old men.”
and
“Higher-BMI older men are actually common.”
The distribution changed at the edge, and the eye remembers the edge.
8. Bottom line
There is real data behind the impression that very large older men are uncommon.
The accurate version is:
Men over 70 with BMI 25-plus are still common, but men with BMI 30-plus are much rarer. Survival selection, late-life weight loss, muscle loss, birth-cohort differences and visibility bias can all contribute.
And because low BMI and unintentional weight loss can also be harmful in later life, “thinner” should not automatically be translated into “healthier.”
A street-corner joke about where the big old men went ends up at a serious conclusion:
In old age, body weight alone tells you surprisingly little about what is happening inside the body.
Sources
- Japan Ministry of Health, Labour and Welfare, 2024 National Health and Nutrition Survey results mhlw.go.jp
- MHLW, 2024 National Health and Nutrition Survey, Part 2: Physical Status mhlw.go.jp
- Sasazuki S, et al. Body mass index and mortality from all causes and major causes in Japanese. J Epidemiol. 2011 pubmed.ncbi.nlm.nih.gov
- Tamakoshi A, et al. BMI and all-cause mortality among Japanese older adults: findings from the Japan Collaborative Cohort Study. Obesity (Silver Spring). 2010 pubmed.ncbi.nlm.nih.gov
- The association of weight change and all-cause mortality in older adults: a systematic review and meta-analysis pubmed.ncbi.nlm.nih.gov
- Donini LM, et al. Definition and diagnostic criteria for sarcopenic obesity: ESPEN and EASO consensus statement. Clin Nutr. 2022 pubmed.ncbi.nlm.nih.gov
- Tarui I, et al. Trends in BMI among elderly Japanese population. Public Health Nutr. 2020 pubmed.ncbi.nlm.nih.gov
- Kosugi T, et al. Association of body indices with mortality in older population: Japan Specific Health Checkups Study. J Am Geriatr Soc. 2025 pubmed.ncbi.nlm.nih.gov
