The influence of occupation on the longevity of Japanese traditional artists

The study investigates whether engaging in strenuous physical activity as part of a lifetime occupation influences longevity. Although abundant epidemiological evidence links regular exercise with reduced mortality, it is unclear if occupations requiring daily vigorous activity confer similar longevity benefits. Japan’s traditional arts offer comparable cohorts differing mainly in occupational physical demand: Kabuki (singing/dancing with intense movements), Noh (exercise), versus largely sedentary Sado (tea ceremony), Rakugo (seated storytelling), and Nagauta (seated musical performance). Given shared contexts—male-dominated professions, urban residence, similar diets, historical records since ~1700—the authors hypothesized that Kabuki and Noh actors would live longer than sedentary counterparts. Shogun and Emperor family members were included as elite comparators presumed to have superior diet and medical care.

Prior studies consistently show that physical activity reduces all-cause and cardiovascular mortality (e.g., Morris et al., 1953; Blair et al., 1989; Sesso et al., 2000; Lear et al., 2017), with benefits observed across income settings and supported by interventions targeting obesity-related risks (Misra and Khurana, 2008; Müller et al., 2016). However, occupational class differences in mortality and higher mortality for manual labor versus non-manual labor have been reported in Europe and Japan (Tanaka et al., 2019), though manual labor does not necessarily imply vigorous activity. Some evidence suggests that very frequent strenuous activity may not yield additional vascular benefits compared with moderate exercise (Armstrong et al., 2015). Other relevant literature addresses sedentary behavior’s harms (de Rezende et al., 2014), stress-related mortality from harmful workplace practices (Goh et al., 2015), and lifestyle factors linked to survival (Rizzuto and Fratiglioni, 2014). Studies on strength training indicate mortality benefits (Stamatakis et al., 2018; Kraschnewski et al., 2016). Historical accounts note potentially unbalanced elite diets (Hanley, 1997) and the past use of leaded face powder in Japan (Yoshinaga, 2015).

Design: Historical cohort analysis of longevity by occupation using archival records. Data sources: Personal-name dictionaries and official webpages for arts events; official sources for Imperial (Emperor) and Shogun family members. Data confirmed as of July 1, 2018. Dataset published (Hayashi and Kezuka, 2020). Population: 699 professional male individuals (546 deceased) born in or after 1700 AD across seven groups: Kabuki, Sado, Rakugo, Nagauta, Noh (preliminary due to sparse data), Shogun family members, and Emperor family members. Only males were analyzed due to male dominance of the professions historically. Inclusion/exclusion: Included individuals with known birth date and, if deceased, death date. Excluded if (1) fewer than two sources confirmed both birth and death dates (exception: Noh actors accepted with one source due to scarcity), (2) death due to battle, suicide, or accident, or death before age 20. Analysis focused on those who reached at least age 20 to reduce infant/child mortality bias. Occupational activity characterization: Kabuki involves singing and dance with intense movements; high prevalence of performance-related injuries reported in prior survey. Sado, Rakugo, and Nagauta are performed seated with minimal occupational physical exertion. Noh includes exercise but had limited reliable data. Workload quantification: For 36 randomly selected Kabuki actors, the number of performance titles per year was counted from association records in 1970, 1990, and 2010; one title typically runs for a month. Ranges were 6–24 (1970), 5–17 (1990), and 9–27 (2010). Performances are ~1 hour; actors also undertake daily lessons and short-term engagements. Statistical analysis: Kaplan–Meier survival analysis with generalized Wilcoxon test to compare survival functions among occupation groups. Discrete-time logistic regression models assessed mortality risk with age as baseline hazard and occupations as predictors, using Shogun and Emperor as reference categories. Birth year controlled in three specifications: Model 0 (no birth-year control), Model 1 (20th-century birth dummy), Model 2 (birth year continuous), Model 3 (birth year plus squared term).

• Sample: 699 male individuals (546 deaths). Kabuki workload ranged annually from 5–27 performance titles depending on year.
• Kaplan–Meier results: Survival curves separated into two groups: (a) Shogun/Emperor family members and (b) all other occupations. Generalized Wilcoxon test showed a significant occupation effect (chi-square = 109.0, p < 0.05, d.f. = 6). Descriptively, Shogun and Emperor family members had shorter lifespans than other groups.
• Table 1 descriptive medians (years): Emperor ~54; Shogun ~54; Kabuki ~71; Noh ~69; Sado ~69; Rakugo ~70; Nagauta ~74 (with corresponding lower/upper 95% bounds reported).
• Discrete-time logistic regression: In the baseline model, Kabuki actors exhibited significantly shorter longevity compared to other traditional artist groups (with Shogun/Emperor as references). After adjusting for birth year (Models 1–3), the Kabuki effect was no longer significant, indicating birth cohort/social-environmental factors accounted for the observed difference. Noh coefficients suggested lifespans similar to sedentary arts (Sado, Rakugo).

The initial hypothesis—that professions involving regular strenuous exercise (Kabuki, Noh) would show longer lifespans—was not supported. Survival analysis showed Kabuki actors had lifespans shorter than or similar to sedentary traditional artists and notably longer than Shogun/Emperor family members. Regression analyses indicated that the apparent shorter longevity of Kabuki actors diminished after controlling for birth year, suggesting cohort effects and improvements in social, sanitary, and medical environments over time play substantial roles. Possible explanations for the lack of occupational exercise advantage include: (1) potential detrimental effects of excessive or highly strenuous occupational activity that may offset benefits observed with moderate exercise; (2) historical exposure to lead-containing white makeup used by Kabuki actors prior to 1934; and (3) potential genetic or familial factors within acting lineages. The authors argue that unique dance movement patterns are unlikely to explain shorter lifespans given evidence of benefits from both endurance and strength training. The shorter lifespans among Shogun and Emperor family members may relate to unbalanced elite diets, high stress, and prolonged sedentary behavior. Overall, the findings indicate that strenuous exercise embedded in occupational duties does not confer the same longevity advantage documented for leisure-time or moderate physical activity, and that broader lifestyle and historical cohort factors substantially influence longevity outcomes.

Daily strenuous exercise as part of an occupation does not prolong longevity among Japanese traditional artists and may be associated with shorter lifespan compared to sedentary arts, though differences are largely explained by birth cohort effects. Elite comparators (Shogun and Emperor families) exhibited the shortest lifespans, likely due to non-physical-activity lifestyle factors. The study highlights the need to determine optimal exercise levels for occupational performers and to consider non-exercise aspects of artistic professions (e.g., speaking, singing, instrument playing) in health outcomes. Future research should include female cohorts, consider retirement effects, examine additional professional activities and exposures (e.g., cosmetics), and assess generalizability across countries and time periods.

• Longevity metric: The study did not estimate life expectancy under a precise demographic framework; analyses excluded individuals who died before age 20 to minimize infant/child mortality bias.
• Sex limitation: Only males were included due to male dominance in these professions historically; female data were lacking.
• Data sourcing and selection: Potential overrepresentation of older, more prominent traditional artists in databases; Noh actor data sometimes confirmed by a single source due to scarcity.
• Comparability: Shogun/Emperor records more likely to include young deaths, potentially biasing comparisons; differing historical exposures (diet, stress, medical care) across groups.
• Cohort effects: Lifespans observed are much shorter than modern Japanese male life expectancy; strong birth-cohort effects necessitate cautious interpretation.
• Unmeasured confounders: Occupational exposures (e.g., leaded makeup prior to 1934), lifestyle factors, and retirement effects were not directly measured.
• Generalizability: Findings may not generalize outside Japan; international patterns of occupational mortality differ.