Body Mass Index and the Risk of Rheumatoid Arthritis
Body Mass Index and the Risk of Rheumatoid Arthritis
With the performance of genome-wide association studies, there have been major advances in our understanding of genetic risk. In addition, there has also been renewed interest in environmental factors, especially lifestyle factors such as smoking, breast feeding, and alcohol intake, which have been identified to have a dose-related association with diseases. As a potential risk factor, several studies have investigated the association of obesity with RA development; however, the relationship remains poorly understood. We therefore conducted a systematic review to comprehensively summarize the current literature on the association of BMI with RA development. We also performed a dose-response meta-analysis to assess whether there was a linear or nonlinear dose-response relationship between BMI and RA risk.
Eleven studies regarding the association between BMI with RA risk showed a positive association between obesity and RA. Compared with non-obese individuals, individuals who were obese had a 24% increased risk for RA development. Individuals who were obese or overweight had a 31% and 15% increased risk for RA in comparison to individuals of normal weight, respectively. The results obtained from a subgroup meta-analysis showed that the RR for RA was high in females compared to mixed populations. One explanation for this phenomenon may be that female sex hormones could modify the effect of obesity on RA risk. The pooled RRs across studies in North American populations were also relatively higher than those of European populations. This may due to most North American studies being designed as cohort studies, which may give relatively higher RRs. For example, the summary RRs from case-control studies were lower than those of cohort studies for obesity versus normal weight (1.22 versus 1.39) and overweight versus normal weight (1.03 versus 1.24), while the RR for obesity versus non-obesity was higher across case-control studies (1.32 versus 1.27). Besides, the results from a sensitivity analysis revealed that the effect of obesity on RA risk in the NHS2 cohort study of Lu et al. was much larger than that observed in the other studies. Lu et al. found that there was not a significant association between BMI and RA when restricting to RA cases diagnosed after 55 years of age, and 83% of RA cases in that study were diagnosed at or before 55 years of age. Hence, the potential explanation for higher risk of BMI on RA risk in the NHS2 study by Lu et al. may be that the age of NHS2 participants was relatively younger.
In addition, all included studies were conducted in North American and European populations. Data for other continents were lacking, and investigators in these regions should pay more attention to the assessment of BMI and RA. Because of the different lifestyles, different living environments, and different economic development between developed countries and developing countries, obesity or increased BMIs may have different effects on the occurrence of disease around the world, especially in Africa. In two cross-sectional studies from African countries, the investigator found that the traditional factors, such as obesity, were not independent risk factors for cardiovascular disease in RA patients when comparing developing black African populations and developed Caucasian populations. These findings suggested that obesity or an increased BMI may have different associations with disease in developing regions or developed regions. In a study by Dessein et al., the authors found BMI, waist circumference, and hip circumference to be relatively lower in African RA individuals than non-RA individuals, and the incidence of Africans with RA decreased overall with abdominal adiposity. These results further support the hypothesis that obesity or BMI may have a diverse influence on RA risk in different regions.
RA can be divided into two major subsets based on the presence of ACPA. Recent genome-wide association studies have shown that significant risk allele frequencies were different between ACPA-positive and ACPA-negative RA patients, which showed a difference in distinct genetic etiologies of those two RA subsets, and provided further support for the need to consider them separately. The association of BMI with disease risk has also been identified to be different in those two subsets. In the study by Wesley et al., findings indicated that obesity is related to ACPA-negative RA development in women, and showed an inverse association between BMI and ACPA-positive RA in men. In the study by Pedersen et al., BMI was also found to be strongly and selectively associated with ACPA-negative RA. However, the study by Lu et al., which contained two cohorts, concluded that the RR of RA was elevated among overweight and obese women both in ACPA-positive and ACPA-negative RA. The results from subgroup analysis by ACPA seropositivity revealed the association of BMI or obesity with RA risk in ACPA-positive RA rather than ACPA-negative RA, suggesting that BMI plays a different role in these two major RA subsets based on the presence of ACPA. However, the great heterogeneity may reflect modification of the relationship of BMI with ACPA-seropositive or ACPA-seronegative RA, and other confounding factors, such as age and gender need to be studied.
Although the mechanism by which obesity or higher BMI could lead to RA remains unclear, there are several potential mechanisms. First is the association between obesity and inflammation. Obesity is often considered a systemic inflammatory condition with increased levels of inflammatory cytokines, including tumor necrosis factor-alpha and interleukin-6. These inflammatory cytokines could promote the inflammatory response of individuals. Leptin, as a pro-inflammatory adipokine, could be secreted excessively in obesity by adipocytes. Previous studies have identified leptin as a potent immune modulator, which could sustain autoreactive cell proliferation and impact inflammation. Both leptin and inflammatory cytokines are implicated in the development of autoimmune diseases. Second, altered sex hormones metabolism in obese subjects may share a similar mechanism in the etiology of RA. Obese individuals have higher levels of estrogens and androgens. Sex hormones have also been shown to play a role in the development of RA, which could be modified by obesity. Besides, the link between obesity and autoimmune diseases could be driven by a genetic variation which could predispose individuals to both conditions.
Of note, there were several limitations in the present study. First, the number of included studies was relatively small. Some of the eligible studies were case-control studies, which were more prone to bias. For example, case-control studies were at a major risk of recall bias, and participants in several studies were asked to report their BMI before interview. The assessment of BMI in subjects from seven included studies was obtained by a self-reported questionnaire, which may influence the accuracy of the data. Second, due to the lack of data provided by each study on the association of BMI categories and RA risk, secondary calculations were required in some cases. These secondary calculations may not exclude the influence of some bias factors on the results. Third, eligible studies only consisted of published data; unpublished data were not identified. This suggests that publication bias cannot be absolutely excluded even though no significant publication bias was detected. Because the high heterogeneity was still observed in subgroup meta-analysis by gender, region, ACPA, and study design, other factors may be the cause of this great heterogeneity, such as rheumatoid factor. However, the role of BMI in RA patients with or without rheumatoid factor could not be determined due to insufficient data. Therefore, it was impossible to completely exclude the influence of inherent confounding factors. In addition, another major limitation is the differences in adjustment for covariates. As shown in Table 1, different covariates were used for multivariable analysis among all included studies. The meta-analysis of included studies inherits the limitation of the original studies. Although most included studies adjusted for potential confounders such as age, gender, smoking, alcohol, and parity, the possibility of residual confounding cannot be ruled out in these studies.
Discussion
With the performance of genome-wide association studies, there have been major advances in our understanding of genetic risk. In addition, there has also been renewed interest in environmental factors, especially lifestyle factors such as smoking, breast feeding, and alcohol intake, which have been identified to have a dose-related association with diseases. As a potential risk factor, several studies have investigated the association of obesity with RA development; however, the relationship remains poorly understood. We therefore conducted a systematic review to comprehensively summarize the current literature on the association of BMI with RA development. We also performed a dose-response meta-analysis to assess whether there was a linear or nonlinear dose-response relationship between BMI and RA risk.
Eleven studies regarding the association between BMI with RA risk showed a positive association between obesity and RA. Compared with non-obese individuals, individuals who were obese had a 24% increased risk for RA development. Individuals who were obese or overweight had a 31% and 15% increased risk for RA in comparison to individuals of normal weight, respectively. The results obtained from a subgroup meta-analysis showed that the RR for RA was high in females compared to mixed populations. One explanation for this phenomenon may be that female sex hormones could modify the effect of obesity on RA risk. The pooled RRs across studies in North American populations were also relatively higher than those of European populations. This may due to most North American studies being designed as cohort studies, which may give relatively higher RRs. For example, the summary RRs from case-control studies were lower than those of cohort studies for obesity versus normal weight (1.22 versus 1.39) and overweight versus normal weight (1.03 versus 1.24), while the RR for obesity versus non-obesity was higher across case-control studies (1.32 versus 1.27). Besides, the results from a sensitivity analysis revealed that the effect of obesity on RA risk in the NHS2 cohort study of Lu et al. was much larger than that observed in the other studies. Lu et al. found that there was not a significant association between BMI and RA when restricting to RA cases diagnosed after 55 years of age, and 83% of RA cases in that study were diagnosed at or before 55 years of age. Hence, the potential explanation for higher risk of BMI on RA risk in the NHS2 study by Lu et al. may be that the age of NHS2 participants was relatively younger.
In addition, all included studies were conducted in North American and European populations. Data for other continents were lacking, and investigators in these regions should pay more attention to the assessment of BMI and RA. Because of the different lifestyles, different living environments, and different economic development between developed countries and developing countries, obesity or increased BMIs may have different effects on the occurrence of disease around the world, especially in Africa. In two cross-sectional studies from African countries, the investigator found that the traditional factors, such as obesity, were not independent risk factors for cardiovascular disease in RA patients when comparing developing black African populations and developed Caucasian populations. These findings suggested that obesity or an increased BMI may have different associations with disease in developing regions or developed regions. In a study by Dessein et al., the authors found BMI, waist circumference, and hip circumference to be relatively lower in African RA individuals than non-RA individuals, and the incidence of Africans with RA decreased overall with abdominal adiposity. These results further support the hypothesis that obesity or BMI may have a diverse influence on RA risk in different regions.
RA can be divided into two major subsets based on the presence of ACPA. Recent genome-wide association studies have shown that significant risk allele frequencies were different between ACPA-positive and ACPA-negative RA patients, which showed a difference in distinct genetic etiologies of those two RA subsets, and provided further support for the need to consider them separately. The association of BMI with disease risk has also been identified to be different in those two subsets. In the study by Wesley et al., findings indicated that obesity is related to ACPA-negative RA development in women, and showed an inverse association between BMI and ACPA-positive RA in men. In the study by Pedersen et al., BMI was also found to be strongly and selectively associated with ACPA-negative RA. However, the study by Lu et al., which contained two cohorts, concluded that the RR of RA was elevated among overweight and obese women both in ACPA-positive and ACPA-negative RA. The results from subgroup analysis by ACPA seropositivity revealed the association of BMI or obesity with RA risk in ACPA-positive RA rather than ACPA-negative RA, suggesting that BMI plays a different role in these two major RA subsets based on the presence of ACPA. However, the great heterogeneity may reflect modification of the relationship of BMI with ACPA-seropositive or ACPA-seronegative RA, and other confounding factors, such as age and gender need to be studied.
Although the mechanism by which obesity or higher BMI could lead to RA remains unclear, there are several potential mechanisms. First is the association between obesity and inflammation. Obesity is often considered a systemic inflammatory condition with increased levels of inflammatory cytokines, including tumor necrosis factor-alpha and interleukin-6. These inflammatory cytokines could promote the inflammatory response of individuals. Leptin, as a pro-inflammatory adipokine, could be secreted excessively in obesity by adipocytes. Previous studies have identified leptin as a potent immune modulator, which could sustain autoreactive cell proliferation and impact inflammation. Both leptin and inflammatory cytokines are implicated in the development of autoimmune diseases. Second, altered sex hormones metabolism in obese subjects may share a similar mechanism in the etiology of RA. Obese individuals have higher levels of estrogens and androgens. Sex hormones have also been shown to play a role in the development of RA, which could be modified by obesity. Besides, the link between obesity and autoimmune diseases could be driven by a genetic variation which could predispose individuals to both conditions.
Of note, there were several limitations in the present study. First, the number of included studies was relatively small. Some of the eligible studies were case-control studies, which were more prone to bias. For example, case-control studies were at a major risk of recall bias, and participants in several studies were asked to report their BMI before interview. The assessment of BMI in subjects from seven included studies was obtained by a self-reported questionnaire, which may influence the accuracy of the data. Second, due to the lack of data provided by each study on the association of BMI categories and RA risk, secondary calculations were required in some cases. These secondary calculations may not exclude the influence of some bias factors on the results. Third, eligible studies only consisted of published data; unpublished data were not identified. This suggests that publication bias cannot be absolutely excluded even though no significant publication bias was detected. Because the high heterogeneity was still observed in subgroup meta-analysis by gender, region, ACPA, and study design, other factors may be the cause of this great heterogeneity, such as rheumatoid factor. However, the role of BMI in RA patients with or without rheumatoid factor could not be determined due to insufficient data. Therefore, it was impossible to completely exclude the influence of inherent confounding factors. In addition, another major limitation is the differences in adjustment for covariates. As shown in Table 1, different covariates were used for multivariable analysis among all included studies. The meta-analysis of included studies inherits the limitation of the original studies. Although most included studies adjusted for potential confounders such as age, gender, smoking, alcohol, and parity, the possibility of residual confounding cannot be ruled out in these studies.
Source...