Background

Nowadays, cardiovascular diseases (CVD) are still considered to be the leading cause of death and disability globally [1]. According to the Global Burden of Disease (GBD) study, the global prevalence of CVD was estimated to have increased by 93% over the past three decades, up to 523 million in 2019 [2], which placed a significant economic burden on health systems and communities. Therefore, exploring the prevention and risk factors of CVD is very important for the management of CVD in future.

It is well known that inflammation plays an important role in the development of CVD [3,4,5]. Rheumatoid arthritis (RA) is a chronic autoimmune inflammatory disease with a global prevalence ratio of about 0.5-1%, which caused damage both joints and extra-articular organs [6, 7]. A number of studies have reported that RA was associated with an increased CVD risk [8, 9]. Løgstrup, et al., demonstrated that patients with RA were more likely to suffer from the risk of heart failure than the general population [10]. Not only that, the risk of CVD was also influenced by lifestyle factors, such as smoking, physical activity and poor diet. Studies suggested that diet may increase the risk of CVD by regulating metabolic pathways and homeostasis [11, 12]. Over the past few years, Mediterranean diet (MD), which has been characterized by high intake of fish, olive oil, fruits, vegetables, whole grains, legumes/nuts, and moderate alcohol consumption, was considered to have an anti-inflammatory effect and one of the best cardiovascular health diets [11].

Recently, a study also found that MD could reduce pain and increase physical function of patients with RA [13], implicating that there might be a combined effect of RA and MD on the risk of CVD. However, to date, there were few studies focusing on the interaction analysis between RA and MD. Thus, our study aims to outline the relationship of coexistence of RA and MD on the risk of CVD for the middle aged and elderly. We believe that the research would be useful reference for reducing the risk of CVD in patients with RA.

Methods

Study population

The study population came from the National Health and Nutritional Examination Survey (NHANES) 2005-2010. NHANES database is a program of studies performed by the Centers for Disease Control and Prevention (CDC) of America using a multistage, probability sampling methods [14], which examined a nationally representative sample of approximately 5,000 people from 15 different counties each year ( https://www.cdc.gov/nchs/nhanes/about_nhanes.htm).

This cross-sectional study extracted 4,750 middle aged and elderly participants (aged ≥ 40 years old) [15] and with complete dietary data from the NHANES database 2005–2010. Exclusion criteria were as following: (1) participants had CVD before suffering from RA (n = 2); (2) participants had unavailable assessment of RA (n = 1,386); (3) participants had extreme total energy intakes of < 500 or > 5000 kcal/day for women, and < 500 or > 8000 kcal/day for men (n = 10). Finally, a total of 3,352 participants were included in this study, and they were divided into CVD group (n = 385) and non-CVD group (n = 2,967) (Fig. 1). The requirement of ethical approval for this was waived by the Institutional Review Board of Hainan Affiliated Hospital of Hainan Medical University, because the data was accessed from NHANES (a publicly available database). All methods were carried out in accordance with relevant guidelines and regulations (declaration of Helsinki). All individuals provided written informed consent before participating in the study.

Fig. 1
figure 1

Participant flowchart of the study

Full size image

Data collection

This study collected some variables [9, 13]: age (years), gender, ethnicity, educational level, marital status, poverty-income ratio (PIR), waist circumference (cm), body mass index (BMI, kg/m2), smoking, drinking, physical activity [metabolic equivalent (MET)·min/week], hypertension, diabetes, dyslipidemia, family history of CVD, RA, MD score, energy (kcal/day/kg), protein (%), carbohydrate (%), total-sugar (%), total-fat (%), C-reactive protein (CRP, mg/dL), uric acid (mg/dL), estimated glomerular filtration rate (eGFR, mL/min/1.73m2).

Waist circumference was classified as abdominal non-obese (< 88 cm for women and < 102 cm for men) and abdominal obese (≥ 88 cm for women and ≥ 102 cm for males) [16]. MET is used to describe the energy consumption when performing a specific activity [17], and physical activity was calculated as weekly energy consumption in this study, weekly energy consumption (MET·min/week) = MET × exercise time of corresponding activity (min)/week. Energy (kcal/day/kg) is defined as total energy (kcal/day) divided by the weight (kg). Hypertension was defined as systolic blood pressure ≥ 140 mmHg or diastolic blood pressure ≥ 90 mmHg or self-reported hypertension. Diabetes was defined as fasting glucose ≥ 7.0mmol/L or self-reported diabetes. Dyslipidemia was determined by total cholesterol ≥ 200 mg/dL or triglyceride ≥ 150 mg/dL or low-density lipoprotein cholesterol ≥ 130 mg/dL or high-density lipoprotein cholesterol ≤ 40 mg/dL. The data of uric acid was collected by timed endpoint method used the Beckman Synchron LX20 or Beckman Coulter UniCel® DxC800, both monitored for absorbance changes at 520 nm. eGFR was calculated by the following equation: 141×min [serum creatinine (Scr)/κ, 1] α×max (Scr/κ, 1)− 1.029 × 0.993age×1.108 (if female) ×1.159 (if black), κ is 0.7 for females and 0.9 for males, α is − 0.329 for females and − 0.411 for males, min indicates the minimum of Scr/κ or 1, and max indicates the maximum of Scr/κ or 1 [18].

Measurements

Assessment of Mediterranean diet

We adopted the alternate Mediterranean Diet Index (aMD) to assess all participants’ adherence to the MD. The aMD scores was determined by assessing the intakes of alcohol, presumed detrimental foods [including red and processed meat], and presumed beneficial foods [seafood, whole grains, legumes, nuts, fruits, vegetables (except potatoes), and the ratio of monounsaturated fish, and the ratio of monounsaturated fatty acids-to-saturated fatty acids] according to the information on 24 h dietary recall questionnaires, which with a total of 9 points [19]. A score of 1 point was assigned to participants who had an intake of presumed beneficial foods higher than the median level, and presumed detrimental foods lower the median level. For alcohol, “1 point” was assigned to participants who had moderate alcohol consumption (10–25 g/day for men and 5–15 g/day for women); “0 points” was assigned for participants who had not a consumption of meeting the aforementioned criteria. The higher the aMD scores, the better of the adherence to the MD. We divided the MD score into two groups based on the median [20]: <3.80 as low adherence to the MD group, and ≥ 3.80 as high adherence to the MD group in this study.

Assessment of rheumatoid arthritis

The diagnosis of RA was based on the self-reported information: “Has a doctor or other health professional ever told you that you had RA?”. The individuals would be diagnosed as having RA if the question were answered as “yes” [21].

Outcomes

The outcome of this study was considered as the incident of CVD. The definition of CVD was based on the self-reported questions: Ever told you had congestive heart failure/ coronary heart disease/angina/angina pectoris/heart attack/stroke. Participants who answered “yes” to the question were defined as having CVD [22].

Statistical analysis

Considered the complex sampling design of the NHANES database, we used a weighted analysis: the sampling weights for interview (WTMEC2YR) and study design variables (SDMVPSU and SDMVSTRA). The mean ± standard error (SE) was used to describe the distributed variables, and comparison between the CVD group and non-CVD group was performed by Student’s t-test. The number of cases and the composition ratio [n (%)] was used to describe the categorical data, and comparison between two groups adopted Rao-Scott Chi-square test. SPSS (version 23.0) and R (version 4.1.2) software was used for statistical analyses.

In the present study, we conducted a weighted univariate logistic regression to screen for confounding factors. Then, weighted multivariate logistic regression was adopted to explore the relationship of RA, MD on the risk of CVD, respectively. Two models were used in this study. Model 1: adjusted age, gender and ethnicity; Model 2: adjusted covariates that were statistically significant in the univariate logistic regression analysis, including age, gender, ethnicity, educational level, PIR, waist circumference, smoking, physical activity, hypertension, diabetes, dyslipidemia, family history of CVD, uric acid, eGFR and energy (Supplementary Table 1). Odds ratio (OR) and 95% confidence interval (CI) were calculated. In addition, we assessed the coexistence effect of RA and MD for the risk of CVD by weighted multivariate regression analysis. The additive interaction between RA and MD on the risk of CVD was evaluated by the relative excess risk due to interaction (RERI), attributable proportion (AP) and the synergy index (SI) [23]. When the 95% CI of RERI or AP were not included 0, or the 95% CI of SI not contained 1, we considered there was an additive interaction. The multiplicative interaction between RA and MD on the risk of CVD was evaluated by OR, when the 95% CI of product-term was not contained 1, we considered there was a multiplicative interaction. Additionally, we also performed subgroup analyses based on age, gender, ethnicity, BMI and family history of CVD, to explore the interaction between RA and MD on the risk of CVD in different population. Statistical significance difference was confirmed by P < 0.05. For the samples with missing data, we adopted.

multiple interpolation methods, and sensitivity analysis was performed on the data before and after interpolation (Supplemental Table 2).

Results

General characteristics of participants

Table 1 displays the general characteristics of participants in CVD group and non-CVD group, and there were some variables with a significant difference between the CVD group and non-CVD group in the distribution of age, gender, educational level, PIR, waist circumference, BMI, smoking, physical activity, hypertension, diabetes, family history of CVD, CRP, uric acid, eGFR, energy. Participants in the CVD group had a lower PIR and MD score, and a higher waist circumference and BMI compared to those who were non-CVD group. Additionally, RA might be more common in the CVD group than non-CVD group. The detailed information was shown in Table 1.

Table 1 General characteristics of included participants
Full size table

The association of RA, MD and the risk of CVD

The association of RA and CVD, MD and CVD were presented in the Table 2. The result of multivariate logistic regression model indicated that RA was associated with an increased risk of CVD (Model 1: OR = 3.98, 95%CI: 2.76–5.73, P < 0.001; Model 2: OR = 2.65, 95%CI: 1.69–4.16, P < 0.001). Compared to those who were high adherence to the MD, those who were low adherence to the MD were positively associated with the risk of CVD [OR of 1.82 (95% CI: 1.13–2.93, P = 0.015)] in Model 1 and OR of 1.67 (95% CI: 1.01–2.77, P = 0.046) in Model 2.

Table 2 The association of RA and CVD, MD and CVD by weighted multivariate logistic regression
Full size table

The coexistence of RA and MD on the risk of CVD

As shown in Table 3, we established the interaction term of RA and MD to assess the additive interaction: RA & low adherence to the MD, non-RA & low adherence to the MD, RA & high adherence to the MD, non-RA & high adherence to the MD. After adjusting confounding factors, we observed an additive interaction of RA and low adherence to the MD on the risk of CVD (RERI = 4.76, 95% CI: 0.52-9.00; AP = 0.74, 95% CI: 0.54–0.95; SI = 8.21, 95% CI: 1.48–45.51), which indicated that there might be a synergistic effect.

Table 3 The additive interaction of RA and MD on the risk of CVD
Full size table

Simultaneously, we also assessed a multiplicative interaction of RA and MD on the risk of CVD (Table 4). After incorporating RA, low adherence to the MD and the product-term of RA & low adherence to the MD into multivariate logistic regression model, we found that the multiplicative interaction between RA and low adherence to the MD on the risk of CVD was statistically significant (OR = 3.63, 95% CI: 1.44–9.15), which also indicated that there might be a synergistic effect. Additionally, the statistic of likelihood ratio test was 8.141 (P = 0.009), suggesting there was an interaction between RA and MD on CVD risk.

Table 4 The multiplicative interaction of RA and low adherence to the MD on the risk of CVD
Full size table

The coexistence of RA and MD on the risk of CVD based on age, gender, ethnicity, BMI and family history of CVD

The performed subgroup analyses based on age, gender, ethnicity, BMI and family history of CVD were conducted. As shown in Table 5, the additive interaction between RA and low adherence to the MD on CVD risk was present in the other subgroup analyses except in those with a family history of CVD. Furthermore, Table 6 also indicated the multiplicative interaction of RA and MD on the risk of CVD based on age, gender, ethnicity, BMI and family history of CVD. Significant multiplicative interaction between depression and low adherence to the MD were found among participants who were non-Hispanic white, aged ≥ 60 years, had not a family history of CVD, and had a BMI of 25–30 kg/m2.

Table 5 The additive interaction of RA and MD on the risk of CVD for different populations
Full size table
Table 6 The multiplicative interaction of RA and MD on the risk of CVD for different populations
Full size table

Discussion

In this study, we verified that RA was associated with an increased risk of CVD, and low adherence to the MD was related to the risk of CVD. An important finding of our study was that the coexistence of RA and low adherence to the MD might be related to the risk of CVD, and there was a synergistic effect.

Previous studies have reported that RA is an inflammatory joint disease, and patients with RA are at higher risk of atherosclerotic CVD, stroke, heart failure, and atrial fibrillation compared to the general population [24, 25]. A review also concluded that treatment of RA may reduce the risk of CVD by reducing chronic inflammation [25]; in other words, the risk of CVD is strongly associated with chronic inflammation in RA disease. Recently, MD has been suggested to be beneficial in preventing and improving CVD [26]. In the study of Critselis E, et al., they found that high adherence to the MD was associated with a lower 10-year risk of fatal and non-fatal CVD [27]. Consistent with established studies [25, 27], our study also indicated the risk of CVD for individuals with independent RA was 2.65 times than non-RA individuals, and the risk of CVD decreased as adherence to the MD increased. As pointed out by Johansson et al., [28] there was an inverse association between MD and RA risk among only men or in seropositive RA. The underlying mechanisms might be that MD’s bioactive compounds decreased the blood pressure, lipids and inflammatory markers.

But, the joint effects of RA and MD on the occurrence of CVD were rarely studied. In the current study, the result indicated that participants with both RA & low adherence to the MD was associated a higher risk of CVD. There was a synergistic effect between RA and low adherence to the MD for the development of CVD. In addition, we found that both additive and multiplicative interactions of RA and low adherence to MD on CVD risk in the analysis of different subgroups were distinguished. The multiplicative interaction was only found among participants who were non-Hispanic white, aged ≥ 60 years, had not a family history of CVD, and had a BMI of 25–30 kg/m2. It is worth noting that neither the additive nor multiplicative interactions of RA and low adherence MD on CVD risk were statistically different in the subgroup with a family history of CVD, which may be related to the small sample size (n = 476).

There were also underlying mechanisms to support the interaction between RA and adherence to the MD on the occurrence of CVD. Mechanisms linking of both RA and the MD on the occurrence of CVD may include shared pathway. Firstly, inflammatory mediators might be a pathway linking of the RA and MD to the risk of CVD [29]. Systemic inflammation is considered as a significant factor in the increased CVD risk among patients with RA [30]. MD patterns rich is in fruits, vegetables, polyunsaturated fatty acids and other nutrients, presenting an anti-inflammatory property, which might feature a protective role for the development of CVD among patients with RA [31]. Secondly, RA is characterized by endothelial dysfunction which tends to be the indicator for CVD occurrence [29]. But the polyphenols in the MD patterns could have a positive effect on endothelial dysfunction [32]. Additionally, oxidative stress is a hallmark of systemic inflammation in RA, and could promote the insulin resistance and further exacerbate the imbalance of reactive oxygen species (ROS) and antioxidants [33], which contributed to the development of CVD. However, MD patterns are associated with a lower level of oxidative stress [34]. When patients with RA have a high adherence to the MD, MD patterns could decrease the level of oxidative stress among patients with RA, which in turn further reduce the risk of CVD occurrence.

This is the first cross-sectional study based on the NHANES database to investigate the interaction between RA and adherence to the MD for the development of CVD. We believe that knowing the interaction of RA & the adherence to the MD might be useful for reducing the risk of CVD in patients with RA. However, some limitations have to be mentioned. Firstly, the assessments of RA, adherence to the MD, and CVD in the NHANES database were based on the self-reported information for participants, which may be subject to recall bias. Nevertheless, several studies have used the self-reported information to examine the factors related to CVD risk [35]. Secondly, we excluded some participants who had unavailable assessment of RA, and we cannot be sure if the excluded participants affect the result. Lastly, this study was a cross-sectional study that could not confirm a causal relationship of RA and adherence to the MD and the risk of CVD. A further study with more focus on the causal relationship is therefore suggested.

Conclusion

RA and MD were associated with CVD occurrence, respectively, and there may be an interaction between RA and MD for the development of CVD. We believe that knowing the interaction of RA and adherence to the MD might be useful for the risk of CVD in patients with RA. However, more prospective clinical studies are needed to further validate our results and explore the possible mechanisms.