Pain is often treated as an inevitable consequence of getting older. A global analysis of more than six million people suggests that picture is too simple. Across 118 countries and territories, self-reported pain generally increased most rapidly before age 55, while different parts of the body followed very different trajectories across the lifespan.
The study, published in Nature Medicine on 5 October 2026, harmonised data from 6,125,459 people across 902 population-based data sources. Participants ranged from age five to more than 100. Researchers mapped self-reported pain at 11 anatomical sites and built global and regional reference curves showing how prevalence changed with age, sex and national development.
Back pain was by far the most common of the site-specific outcomes, affecting roughly 40% of participants across the pooled data, while facial pain was reported by around 2%. Women reported more pain than men at every anatomical site examined, with female-to-male risk ratios ranging from 1.08 to 1.83.
But one of the clearest findings concerned timing. The burden did not simply rise at a constant rate into old age. Across all sites, prevalence increased most steeply before age 55. Some pain types peaked surprisingly early, while musculoskeletal pain continued climbing much later.
Pain did not follow one universal ageing curve
Matt Fillingim, Christophe Tanguay-Sabourin, Lindsay Neuert and colleagues assembled individual-level data collected between 1990 and 2025 from 138 population-based study programmes. The final dataset represented every inhabited continent and included 55% female participants.
The team harmonised study-specific questions and body maps into 11 standard anatomical categories: head, face, neck or shoulder, foot or ankle, hand or wrist, elbow, chest, back, stomach or abdomen, hip and knee. Where enough information was available, they also constructed a measure of any bodily pain.
This approach matters because pain is unusually difficult to compare across populations. There is no objective biomarker that can independently confirm how much pain someone experiences. Health systems also differ in access, diagnosis and coding, meaning estimates based only on diagnosed conditions may miss people whose pain never reaches a clinical pathway.
By using self-reported anatomical pain, the researchers created a complementary framework that can be compared across countries even when healthcare access differs substantially.
Headache peaked at 34 while abdominal pain peaked at 23
Seven of the 11 anatomical sites showed an inverted-U pattern, meaning prevalence rose, reached a peak and then declined rather than increasing indefinitely with age.
The timing varied markedly by body site. Headache prevalence peaked at about age 34 in the global reference trajectory, while stomach or abdominal pain peaked at about 23. Neck or shoulder pain peaked around age 62, elbow pain around 59 and facial pain around 45.
Musculoskeletal pain behaved differently. Back, hip and knee pain showed progressive nonlinear increases through adulthood, with the most rapid gains generally occurring between ages 20 and 55 and peaks at or beyond age 75.
These differences challenge the idea of a single phenomenon called age-related pain. A 30-year-old with recurrent headaches and an 80-year-old with hip pain occupy very different epidemiological trajectories, even though both outcomes are grouped under the broad label of pain.
Women reported more pain at every site studied
Sex differences were remarkably consistent. Women had higher reported prevalence than men across all 11 anatomical sites, with risk ratios ranging from 1.08 to 1.83.
The size of the gap varied by site, but the direction did not. This does not establish why the difference exists. Biological factors, exposure patterns, social conditions, healthcare experiences and differences in pain reporting may all contribute, and a global observational analysis cannot separate those mechanisms cleanly.
What the scale of the dataset does provide is a strong descriptive benchmark. The female excess was not confined to one region, one age group or one type of pain. It appeared across the full set of anatomical outcomes examined.
That is important for research and healthcare because a reference curve can reveal when a local population or clinical group differs from the global expectation. The authors have made the trajectories available through an open benchmarking tool so external datasets can be positioned against the global norms.
Late-life pain was much higher in lower-development countries
The study also produced a result that runs against some condition-based global estimates. From around age 40, any bodily pain, joint pain and back pain rose more steeply in countries with lower Human Development Index values.
By age 80 and older, the estimated prevalence of any bodily pain was 30.6 percentage points higher in low-HDI settings than in high-HDI settings, with a 95% confidence interval from 26.9 to 34.3 percentage points.
For joint pain, the corresponding difference was 21.0 percentage points, while back pain differed by 20.2 percentage points. Low back pain in the lowest-development settings approached twice the prevalence seen in the highest-development settings at the oldest ages.
These gaps emerged mainly after midlife. That pattern suggests that the conditions under which people work, age, access care and accumulate disease may matter greatly for how pain develops later in life.
Sub-Saharan Africa followed a different pattern
Regional differences were not limited to the overall amount of pain. In some cases, the shape of the lifespan trajectory changed.
Headache showed particularly strong geographic variation. North America and western Europe followed the global pattern of an early-adulthood peak followed by decline. East and Southeast Asia had lower and flatter prevalence across the lifespan. In sub-Saharan Africa, headache instead showed a modest increase from adolescence into later life.
Late-life pain was also particularly elevated in several regions that have historically been underrepresented in global surveillance, including sub-Saharan Africa, central and southern Asia, and north Africa and west Asia.
For South Africa and neighbouring countries, this makes the study more relevant than a global average alone would suggest. The findings indicate that pain trajectories observed in wealthier countries cannot simply be assumed to describe populations in lower-development settings.
Three familiar risk factors explained only part of the burden
The researchers also estimated how much site-specific pain could be attributed to three modifiable risk factors: smoking, obesity and low household income.
Across the global dataset, the combined population-attributable fraction was 18.3%. That figure varied substantially by region, from 12.6% in sub-Saharan Africa to 27.1% in eastern Europe.
The lower attributable fraction in sub-Saharan Africa is important. It does not mean smoking, obesity or socioeconomic conditions are unimportant. Rather, these three measured factors explained a smaller share of the observed pain burden there, leaving more of the regional excess unexplained by the model.
Potential contributors could include occupational exposures, untreated injuries, infectious or inflammatory disease, healthcare access, comorbidity and other social or environmental factors. The current study cannot determine which of these account for the gap, and the authors specifically argue that pain determinants in lower-HDI settings remain poorly characterised.
Prevention may need to begin well before old age
If pain accumulated primarily after retirement age, prevention strategies focused on older adults would make intuitive sense. The new trajectories suggest that this would miss a critical part of the burden.
Across anatomical sites, prevalence rose most steeply before 55. For several pain types, the peak occurred in early or middle adulthood. That means working-age populations are not merely on the way toward the main pain burden of later life. For some conditions, they are already passing through the period of fastest increase or highest prevalence.
This has implications for workplace health, primary care and public-health surveillance. Earlier prevention of musculoskeletal strain, obesity and smoking may matter, but the unexplained burden in lower-development settings also shows why generic lifestyle advice cannot substitute for understanding local causes.
The study does not test any intervention, so it cannot show that starting a particular programme before 55 will prevent later pain. Its contribution is to identify when and where the burden changes most sharply, giving future prevention research a more precise target.
The curves are population benchmarks, not personal forecasts
The enormous sample size can make the results sound more predictive for individuals than they really are. These are population-level reference trajectories. They do not tell a particular person when pain will begin, peak or resolve.
Most contributing data were cross-sectional. In other words, researchers generally compared people of different ages rather than following the same individuals from youth into old age. A lower headache prevalence among older adults therefore does not prove that a person’s headaches will necessarily improve as they age.
Cohort effects can also influence the curves. People born in different decades experienced different healthcare systems, occupations, lifestyles and environmental exposures. Those differences may partly contribute to apparent age patterns.
Self-reported pain is another necessary limitation. It captures an experience that cannot be reduced to a laboratory measurement, but survey wording, cultural interpretation and reporting behaviour can vary. The researchers harmonised questions across hundreds of sources, yet perfect equivalence between every survey is impossible.
Geographic coverage was also uneven. Western Europe contributed many more data sources per country than sub-Saharan Africa, even though the latter produced some of the most important departures from high-income patterns. The study therefore improves representation without eliminating longstanding global data gaps.
A common language for a problem that has been difficult to compare
The most useful output may ultimately be the reference framework itself. Pain research has long struggled with fragmented definitions, different body sites, inconsistent survey questions and unequal access to clinical diagnosis. Harmonising more than six million self-reports does not solve those problems, but it creates a shared baseline.
The resulting picture is more complex than the assumption that pain simply gets worse with age. Back, hip and knee pain do tend to climb into later life. Headache and abdominal pain peak much earlier. Women report more pain across every site studied. And in lower-development settings, the late-life burden can be dramatically higher than in wealthier countries.
For health systems, that shifts the question from how much pain exists to when it emerges, where it accumulates and which populations diverge from the expected trajectory. The answers suggest that pain prevention cannot be reserved for old age, and that global strategies will need to account for substantial regional differences rather than treating one lifespan curve as universal.
Source Information
Study: Fillingim, M., Tanguay-Sabourin, C., Neuert, L. et al. “Global and regional reference curves for pain across the lifespan in 6.1 million individuals in 118 countries.”
Journal: Nature Medicine.
Published: 5 October 2026.
DOI: 10.1038/s41591-026-04696-w
Study design: Harmonised individual-level analysis of self-reported pain from 6,125,459 participants across 902 population-based data sources and 138 study programmes in 118 countries and territories, collected between 1990 and 2025.
Participants: Ages five to more than 100 years; 55% female.
Primary outcomes: Self-reported pain at 11 anatomical sites, with additional analyses of any bodily pain, high-intensity pain and generalised pain where data permitted.
