Multimorbidity, a term commonly used to describe the presence of two or more chronic physical conditions (1), has been widely acknowledged as a public health challenge. Despite the inconsistency in measurement, the prevalence of multimorbidity is ranging from 30% to 95% across age groups and countries (2, 3). People with multimorbidity suffer more pain, mobility limitations, and higher mortality (4, 5), and secondary prevention of a wide range of chronic diseases deserves attention in this population. To slow chronic disease progression and reduce mortality, regular physical activity (PA) has been recommended by the WHO in the multifaceted care of individuals with chronic disease or multimorbidity (6, 7). At the macro level, PA helps to reduce the disease burden and loss of economic output associated with the treatment of chronic disease complications, especially in low- and middle-income countries (LMICs) (8).

Evidence exists about the different levels of PA in people living with multimorbidity. For example, a study conducted on 46 LMICs revealed that people with chronic conditions and multimorbidity were significantly less physically active (9). Another study conducted on 96,706 United Kingdom Biobank participants (aged 40 years old or older) suggested that participants with chronic disease undertook 9% or 61 min less moderate activity and 11% or 3 min less vigorous activity per week than individuals without chronic disease (10). However, limited studies clarify whether and to what extent multimorbidity is associated with leisure-time physical activity (LTPA), the most important domain of PA (11). Moreover, multimorbidity of chronic disease may be present in multiple system groups (e.g., cardiorespiratory system (12) and psychiatric system (13)), and therefore usually refers to systematic multimorbidity (i.e., intersystem multimorbidity and intrasystem multimorbidity). Understanding the level of LTPA in people living with intersystem or intrasystem multimorbidity can help clinicians for better guidance on rehabilitation exercises to reduce the risk of complications, disability, and death, but such evidence is currently lacking.

Numerous studies have explored racial and ethnic disparities in PA worldwide (14, 15) and found racial/ethnic minorities are especially unlikely to engage in PA and tend to have poorer health outcomes (16). As a systematic review indicated, ethnic differences should also be considered in the relationship between chronic disease and PA (17). However, such studies are scarce, especially for LTPA.

To help fill these gaps, this study aimed to estimate the association between intersystem and intrasystem chronic disease multimorbidity and LTPA by ethnicities based on baseline data from the China Multi-Ethnic Cohort (CMEC) (18). Our study can help clinicians to better understand the levels of LTPA among different multimorbidity groups within Han and ethnic minorities groups, for better management of chronic disease and multimorbidity.

The analytic sample consisted of 76,084 participants aged 30–79 years old. Overall, the mean (SD) age was 51.2 (11.4) years old, 29,392 (38.6%) were men, and 32,059 (42.1%) were minorities. In brief, more than half (62.9%) of the participants engaged in a low level of LTPA, and the minority population (74.7%) was significantly larger than the Han ethnic population (54.3%). For minorities, participants engaged in low LTPA were more likely to be women who were younger, with low annual family income, low educational levels, current smoking, cohabiting, often drinking, high levels of non-LTPA, and long sleep durations. However, for the Han ethnic population, low LTPA tended to appear in men with the aforementioned characteristics (Table 1).

The prevalence of low LTPA in participants with each type of multimorbidity by ethnic group is shown in Figure 1. In brief, the prevalence of low LTPA among minorities with multimorbidity was significantly higher than that of the Han population. The ORs and 95% CIs for the associations of low LTPA with 17 chronic diseases are shown in Supplementary Table 2.

The results of the multivariable logistic regression analysis assessing the associations of multimorbidity, intrasystem multimorbidity, and intersystem multimorbidity with low LTPA are presented in Figure 2. In the overall sample, we observed that respiratory system multimorbidity (OR = 1.10, 95% CI = 1.03–1.18) and respiratory-metabolic system multimorbidity (OR = 1.14, 95% CI = 1.05–1.24) were associated with an increased risk of low LTPA in the adjusted models, while significantly negative associations were observed for multimorbidity (OR = 0.92, 95% CI = 0.89–0.95), digestive system multimorbidity (OR = 0.86, 95% CI = 0.82–0.91), circulatory-digestive system multimorbidity (OR = 0.87, 95% CI = 0.83–0.92), circulatory-metabolic system multimorbidity (OR = 0.92, 95% CI = 0.88–0.96), and digestive-metabolic system multimorbidity (OR = 0.91, 95% CI = 0.87–0.95). Notably, when we considered total PA (Supplementary Table 3), we found that all types of multimorbidity were positively associated with low PA. The intensity matrix suggested some types of multimorbidity (i.e., respiratory and respiratory-digestive) were closely associated (Supplementary Figure 2). The overall interaction between multimorbidity and ethnic group was not significant (OR = 0.95, 95%CI = 0.89–1.02).

When the analysis was stratified by the ethnic group, we found some interesting results (Figure 2). Minority population suffered from multimorbidity tended to have adequate LTPA (OR = 0.86, 95% CI = 0.82–0.91), while a significant association was not found among the Han population (OR = 0.96, 95% CI = 0.92–1.01). In addition, for intrasystem multimorbidity and intersystem multimorbidity, all significant associations with low LTPA were negative in the minority population but were mainly positive in the Han population. For both the minority and Han populations, digestive system multimorbidity and digestive-metabolic system multimorbidity had negative, significant associations with low LTPA. For the Han population, the association of circulatory-respiratory system multimorbidity (OR = 1.17, 95% CI = 1.04–1.31) with low LTPA was stronger than those of circulatory system multimorbidity (OR = 1.12, 95% CI = 1.01–1.25) and respiratory system multimorbidity (OR = 1.14, 95% CI = 1.04–1.25).

The sensitivity analyzes suggested similar effect estimates for low LTPA (Supplementary Table 4). Compared to the corresponding OR in Figure 2, for the Han population, the OR in the association between low LTPA and intrasystem multimorbidity of circulatory, respiratory, and circulatory-respiratory system multimorbidity increased; for minorities, the OR in the associations of low LTPA with circulatory system multimorbidity, circulatory-metabolic system multimorbidity, and digestive-metabolic system multimorbidity decreased.

This study examined the associations of multimorbidity, intrasystem multimorbidity, and intersystem multimorbidity with low LTPA among Han and minority populations based on a large-scale, multiethnic cohort in China. Our findings revealed people with chronic disease multimorbidity were inclined to engage in more LTPA in the overall sample, and the association in minority populations was stronger than that in the Han population. In addition, the ethnic differences in the associations between multimorbidity and low LTPA were not found as the interaction between multimorbidity and ethnic group was not significant. These findings could enable clinicians and policymakers to carry out precise PA interventions and treatments for those with particular chronic disease multimorbidity.

To date, many studies have demonstrated that people with chronic diseases or multimorbidity tend to engage in less PA (21, 32–34). However, evidence about the association of chronic disease multimorbidity with LTPA has been relatively scarce. Considering that LTPA is largely modifiable, our study focused on LTPA and observed people with multimorbidity tend to engage in more LTPA, which could provide important evidence for chronic disease multimorbidity control. Notably, our finding is inconsistent with those of the studies aforementioned (21, 32–34). The most likely reason could be that our outcome variable is not the same as theirs, and we only focused on one domain of PA. As shown in Supplementary Table 3, we found that every multimorbidity was positively associated with low PA when considering PA related to occupation, household, transport, and leisure time as the outcome variable. In addition, a person living with multimorbidity may be recommended by clinicians to intentionally engage in more LTPA. Similar to our results, a negative association between chronic diseases and the different domains of PA was found in a study in Austria involving 8,251 participants (35). Our findings proved again that it is necessary to explore the relationships between different domains of PA and multimorbidity.

There are some interesting findings in the association between multimorbidity and low LTPA by the ethnic group. Among minority populations with multimorbidity, digestive system multimorbidity, digestive-metabolic, circulatory-metabolic, or circulatory-digestive system multimorbidity took part in more LTPA. As shown in Supplementary Table 4, the ORs for most multimorbidity decreased compared with Figure 2 among minorities, suggesting that minorities with multimorbidity were likely to engage in adequate LTPA when they have learned about their body status. In addition, people might not expect digestive-metabolic system disorders to limit activity (10). Conversely, LTPA can contribute to the therapy of chronic diseases, leading to a higher level of LTPA in the population with chronic diseases (36). For the Han population with respiratory system multimorbidity or circulatory system multimorbidity, their LTPA levels were lower. As shown in Supplementary Table 4, the ORs for all types of multimorbidity increased compared with Figure 2 except for respiratory-metabolic system multimorbidity among the Han population. One reason could be that the disease might be directly responsible for lower activity levels due to a reduced exercise capacity by influencing cardiopulmonary function (37). A UK Biobank study of 52,556 participants reported that people with the worst cardiometabolic diseases performed approximately half of moderate to vigorous PA on a daily basis compared to healthy individuals (38). Alternatively, physical inactivity leads to an increased risk of these diseases, and individuals might have been habitually less active than their healthy counterparts for some time to develop the condition (10). Participants with lower LTPA had a greater likelihood of developing chronic diseases, such as metabolic syndrome (39). Moreover, a cohort study conducted for 3 years found that the number of patients with metabolic syndrome having sufficient PA did not significantly increase despite advice to increase PA (40). Multimorbidity can contribute to dementia and cognitive decline (41), which could influence LTPA. For digestive system multimorbidity and digestive-metabolic system multimorbidity among the Han population, their associations with LTPA could be attributed to the same reasons aforementioned in the minority population. The association between chronic conditions or multimorbidity and LTPA might differ in ethnicities due to different disease profiles (42), suboptimal treatment of chronic conditions (43, 44), differences in knowledge regarding the benefits of PA (45), or other environmental factors, such as work conditions (46). Han populations with multimorbidity tended to report low LTPA in this study, contradicting our current knowledge that ethnic minorities engage in less healthy behaviors than the Han people. More future studies are warranted to verify such findings.

We observed that intersystem multimorbidity has a possible synergistic effect on LTPA. For example, among the overall sample, the ORs for low LTPA with respiratory system multimorbidity and metabolic system multimorbidity were 1.10 and 0.99, respectively. However, the effect values increased to 1.14 in the association between low LTPA and respiratory-metabolic system multimorbidity. Consistent synergistic effects among circulatory-respiratory system multimorbidity, respiratory-metabolic system, and low LTPA were also observed among the Han population. The effect has some biological plausibility, with the consideration of several pathways. First, shortness of breath caused by chronic respiratory disease might be worse due to the additional influence of chronic circulatory disease. It seems to be difficult for patients to engage in sufficient LTPA owing to breathing problems (47). Second, the connectedness of several chronic diseases might increase symptom severity, especially pain perception, which could in turn explain why patients with chronic respiratory disease and other chronic diseases avoid LTPA (48). Finally, there is evidence that diseases belonging to common patterns of multimorbidity can interact, curtailing compensatory mechanisms and resulting in more severe physical inactivity (37).

As is known to all, limitations in physical and cognitive function due to multimorbidity decisively reduce people’s levels of PA, which can in turn increase multimorbidity (19, 20). However, our study found people suffering from multimorbidity may engage in adequate LTPA, which suggested that clinicians should not blindly advise patients with multimorbidity to do physical exercise in their spare time.

Several limitations to this study should be considered. First, eight chronic diseases were self-reported based on the questionnaire, and the results could have been affected by information bias. Second, PA could have been influenced by many factors. Although it was impossible to adjust for all of the potential confounding factors, variables such as green space, the built environment, and cognitive function (41, 49, 50) should have been considered. Third, our research only focused on seventeen chronic diseases with a high prevalence in southwest China and excluded people with special health statuses (e.g., infectious diseases, pregnancy, and injury) that might severely affect PA; other diseases affecting PA were not considered either. Fourth, since this study was a cross-sectional study, caution should be taken in making causal interpretations between LTPA and multimorbidity. Nonetheless, our study has notable strengths. Our findings clarified multimorbidity into intrasystem multimorbidity and intersystem multimorbidity, which helps us comprehensively understand the associations of LTPA with specific multimorbidity subgroups.

This large-scale epidemiologic study provides an improved understanding of the impact of multimorbidity on LTPA in Han and minority populations in southwest China. Our findings suggest that LTPA-tailored interventions should be designed for specific ethnic groups according to a different type of multimorbidity. Future large, prospective studies are required to further determine the temporality of the associations observed in this investigation and better explain whether changes in the nature of the dependent variable produce different results.

The original contributions presented in this study are included in the article/Supplementary material, further inquiries can be directed to the corresponding authors.

The studies involving human participants were reviewed and approved by the Ethics Committee of Sichuan University (K2016038). The patients/participants provided their written informed consent to participate in this study.

YL, QN, and JZ: conceptualization. JL, RH, and QN: data curation. YL, XL, and BY: formal analysis and writing—original draft. YL, QN, JL, and RH: investigation. YL, BY, XL, and JZ: methodology. QN: project administration. JZ, YL, XL, and BY: writing—review & editing. All authors have read and agreed to the published version of the manuscript.