Two of the largest burdens on older adult populations are pain and central adiposity. These factors are associated with increased health risks and healthcare costs as well as decreased quality of life (1). Pain is associated with reduced physical function, socialization, and sleep quality (2). In addition, it commonly co-occurs with fatigue (3). Centrally accumulated body fat is considered a component of metabolic syndrome and is specifically linked to insulin resistance, coronary heart disease, cerebrovascular disease, and all-cause mortality (4). Studies indicate that central obesity is linked to several pain outcomes in the older adult population (5–8).

In a study of 7,601 adults aged 65 years and older, 52.9% reported being bothered by pain in the last month and 74.9% of those adults reported multiple sites of pain (9). The prevalence of pain reporting was highest among those with obesity (9). A cross-sectional study of 407 older adults found that those with central obesity had nearly double the odds of chronic pain prevalence as well as increased pain severity and frequency compared with those without central adiposity (5). While the temporality of the association could not be determined, it has provided a foundation for additional research (5).

Central adiposity has not yet been studied as an exposure in a longitudinal study in older adults observing pain outcomes. In this study, waist circumference will serve as a proxy measure of central adiposity, as it has been established as an appropriate measure for central adipose tissue in women and men (10). Understanding how waist circumference affects pain outcomes over time can help focus public health efforts. The present study aimed to explore how central obesity was associated with multisite pain experiences in older adults. We hypothesize that larger waist circumferences are associated with increased rates of multisite pain.

Table 1 describes baseline demographics, physical and mental health measures, and pain characteristics of the NHATS cohort used in this study. Study sample weights were accounted for in the weighted analyses. The study population was predominantly female (56.2%) and non-Hispanic whites (69.3%). Bothersome pain in the last month was reported by 52.7% of individuals, and 47.8% of participants met the definition of multisite pain. Of those that did not have multisite pain, 30.0% of participants did not have any pain sites and 22.2% had only one pain site. This cohort at baseline was 27.3% obese and 41.6% of participants had waist circumferences that were considered high risk based on their gender. Arthritis was the most common comorbidity (55.4%).

Table 2 describes the bivariate association of waist circumference risk with all covariates assessed in this study. The low waist circumference risk is <109 cm in men or <98 cm in women. At baseline, those with high-risk waist circumferences had significantly more multisite pain than those with low-risk waist circumferences (p < .0001). Those with a high-risk waist circumference were also significantly more likely to be bothered by pain, women, obese people, and non-Hispanic whites. Further, having an education of high school or less, probable major depression, and a 2+ multimorbidity count were all associated with high-risk waist circumference. Smoking history, osteoporosis, dementia, and having a fractured hip after age of 50 years were not significantly associated with a higher waist circumference risk level.

Missing values for pain responses were examined over time and by the WC risk group to inform any potential bias. At any given round, no more than 10.1% of the sample was missing a pain response for a group. The breakdown of missing pain response data by the WC risk group at each round is presented in Table 3.

The final model adjusted for cluster sampling and the following covariates: gender, age, race, education, probable major depression, arthritis, and multimorbidity count. In this model, having a high-risk waist circumference increased the rate of having multisite pain by 10.6% [rate ratio (RR) 1.106, 95% CI: 1.066–1.147], adjusting for all covariates (Table 4). The interaction term between waist circumference risk level and gender was not significant in the model (p = .786) and therefore, gender did not modify the association between waist circumference and having multisite pain on the multiplicative scale. Having arthritis was a confounder of the relationship between waist circumference risk level and having multisite pain.

Body mass index was not surprisingly found to be collinear with waist circumference, and thus we did not include BMI or obesity, determined by BMI, in the model. However, we did test BMI as a dichotomous variable (BMI <30 or ≥30) in the model and found that while it did confound the relationship between waist circumference risk and having multisite pain, waist circumference was still significant at p < .0001 when in the fully adjusted model (data not reported).

The findings from this cohort study indicate that there is a significant relationship between waist circumference risk level and the rate of having multisite pain. We found that those with a high-risk waist circumference had a 11% increased rate of having multisite pain compared to those with low-risk waist circumferences, even after adjusting for gender, age, race, education, probable major depression, arthritis, and multimorbidity count. Notably, women in this study were more likely than men to have multisite pain in the adjusted model. These results are consistent with findings from previous cohort and cross-sectional studies that have assessed the relationship between central adiposity and pain in the older adult population (5–8, 21).

While BMI was not included in the final model due to collinearity with waist circumference, it is interesting that when tested in the fully adjusted model, waist circumference was still significant when BMI was in the model. A prior study assessed waist circumference and BMI together in a model assessing pain outcomes but found that the association of BMI and pain outcomes was stronger than that of waist circumference and BMI (7). However this study does not address the claim of other researchers that including waist circumference and BMI in a model together could give a better description of fat distribution, namely, central adiposity (13).

Further research needs to be done on the relationship between waist circumference and BMI in the older adult population to determine the most appropriate way to model central adiposity. Our study is strengthened by employing 8 years of data from the NHATS longitudinal cohort. The waist circumference measure was administered using a rigorous and validated research protocol, which instills confidence in the measure. The pain was ascertained by a researcher during an in-person interview that presented very clear and concrete questions about pain location. If there was any recall bias about pain, it would probably be a non-differential overestimation of multisite pain. Since researchers asked participants about pain in several listed locations, a participant might have been more likely to say “yes” to a location of pain than if they were asked to list where pain bothers them. This study has several limitations. NHATS does not have a comprehensive pain assessment which makes it difficult to determine the extent to which an individual is in pain. This study is also limited in that many variables are self-reported, such as height, weight, and comorbidities. For instance, BMI should be viewed with caution because height and weight were self-reported and may be underestimated. Ideally, future studies would consider additional confounders, such as daily physical activity and also pain medications used by participants (13). Furthermore, while only a small percentage of participants were missing pain responses at each round, findings may be biased as a result. In most rounds, the proportion of missing responses for each group is relatively similar. However, there are several rounds where the low-risk WC group is missing a slightly greater percentage of pain responses, and the computed rate ratios may be biased.

The use of a multimorbidity count is also a limitation of this study, as it possibly oversimplifies the impacts of certain comorbidities on multisite pain. Patients with different comorbidities but with an equal number of them are categorized in the same multimorbidity count group. Except for depressive symptoms, arthritis, osteoporosis, and dementia or Alzheimer's disease, comorbidities (e.g., heart attack, diabetes, lung disease, stroke, and cancer) are examined as an aggregate count. Therefore, we cannot comment on the relationship between specific comorbidities and multisite pain. Because of the limited sample size, we chose to employ this simplified method. However, we did choose to separately examine comorbidities that have been examined by or found to be associated with pain in previous studies (9, 22, 23). For example, Vennu et al. found that obesity was associated with multisite pain in the lower extremities in a population with arthritis (22). In terms of waist circumference, Costa et al. (23) reported an adjusted odds ratio (OR) of 2.03 (1.57–2.63) for waist circumference and multisite chronic pain while adjusting for depression .

One of the challenges of pain research is that pain can be measured in a number of different ways (e.g., by intensity, severity, duration, and location) (24). The pain PROMIS measures, created by the National Institute of Health to target specific pain-related domains, are changing this but currently, there is not one common scale that is used across most epidemiological research. Heim et al. (7) performed a prospective study on a cohort of 2,000 older adults enrolled in the Longitudinal Aging Study Amsterdam. This study assessed pain using a modified Nottingham Health Profile that seeks to understand overall body pain, and they found two- to three-fold increased odds of incident pain after 6 years for those in the highest waist circumference quartile vs. the lowest (7). This association is much stronger than the one found here, which can be attributed to study differences, but also to the measurement of a different aspect of pain.

Measuring methods aside, it will become increasingly important to mitigate pain by all means possible for populations as they age. This study supports that central adiposity, as measured by waist circumference, is associated with an increased rate of multisite pain in older adults. While more research is needed, reducing waist circumference may prove beneficial in reducing the burden of multisite pain. Obesity remains a debated topic in older adults because the changes to the body during aging have made it harder to identify if obesity is even clinically relevant in this population (8, 21). It seems that central adiposity and loss of fat-free mass are more clinically relevant for older adults (21). While weight circumference is not a perfect measure of central adiposity, it might be more clinically relevant to assess an older adults' waist circumference risk vs. their scale weight. Moreover, finding more evidence to support age-appropriate waist circumference cut points in older adults will help researchers establish clinically significant levels that will make screening and identifying people at risk easier.

Publicly available datasets were analyzed in this study. This data can be found at: https://nhats.org.

The studies involving human participants were reviewed and approved by the Columbia Irving Medical Center Institutional Review Board (AAA57372). Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

TM conceptualized the study and provided supervision of the study. CD and TM conducted the data analysis, the methodology, the data visualization, and wrote the original draft. ST and KS provided an in-depth review of the manuscript and assistance in the scientific writing, as well as contribution to data interpretation. All authors contributed to the article and approved the submitted version.

The National Center for Injury Prevention and Control, Centers for Disease Control and Prevention to Columbia University (R49 CE002096-01). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the centers for Disease Control and Prevention. U.S. Department of Health and Human Services, National Institutes of Health, National Institute on Aging (NIA U01AG032947).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.