The relationship between disease activity and quality of life in rheumatoid arthritis patients: a network analysis

Introduction

Rheumatoid arthritis (RA) is one of the most common chronic inflammatory diseases, which can cause joint damage and even disability (Smolen, Aletaha & McInnes, 2016). In recent years, the global prevalence of RA has been continuously increasing. In China, the prevalence of RA is approximately 0.28% (Finckh et al., 2022). As RA cannot be completely cured, “disease improvement” has become an important prognostic indicator, and quality of life (QoL) is one of the most important indicators (Smolen et al., 2020).

In the pathology of RA, cytokines such as IL-1, TNF-alpha, and IL-6 are secreted, which can enter the brain and disrupt the metabolism of monoamine neurotransmitters, affecting normal neurological and cognitive functions and promoting depressive symptoms (Liu, Ho, & Mak, 2012; Pan et al., 2011). RA and depressive symptoms have been found to interact in a large longitudinal study, and RA patients with depressive symptoms have lower QoL scores (Lu et al., 2016; Matcham et al., 2014; Tański, Szalonka, & Tomasiewicz, 2022). Network analysis has shifted from studying individual groups to comparing the network structure of different subgroups (Van Borkulo et al., 2015). In a study of obesity patients, differences were found between the QoL network structures of patients with low and high physical performance (Dalle Grave et al., 2020). However, no study has explored whether there are differences in the network structure of QoL between RA patients with and without depressive symptoms.

According to EULAR guidelines, “remission” is the primary treatment target for RA patients. When “remission” is not feasible, “low disease activity” is accepted to inhibit disease progression (Smolen et al., 2023). However, the remission rate for RA patients is relatively low, approximately 5%–45% (Ajeganova & Huizinga, 2017). A study in Ireland indicates that RA patients who achieve clinical remission or low disease activity have significantly higher QoL scores than those with moderate or high disease activity (Murphy et al., 2024). The same conclusion has been validated in the Portuguese population (Rosa-Gonçalves, Bernardes & Costa, 2018). These studies assess the QoL as a single, unified structure. In fact, the QoL is made up of several domains (Matcham et al., 2014). The complex relationship between disease activity and these domains is still unclear.

In addition, it may be difficult to accurately assess the relationship between disease activity and QoL because of the strong correlation between QoL domains (Koh et al., 2006). Most previous studies have considered QoL as a single overall construct and have not considered correlations between QoL domains (Bai et al., 2023; Zhang et al., 2023). This makes it difficult to assess the impact of disease activity on different domains of QoL in RA patients based on previous research. Network analysis can control the correlation impact between QoL domains and identify the related domains to disease activity (Bai et al., 2023; Takahashi et al., 2023).

This study aims to (1) compare whether there is a difference in the structure of the QoL network between RA patients with and without depressive symptoms and (2) explore the relationship between disease activity and QoL in RA patients.

Material and methods

Results

A total of 424 RA patients were included in this study. Table 1 shows the demographic and disease characteristics of RA patients without (n = 235) and with (n = 189) depressive symptoms. The participants had a mean age of 60.22 ± 12.15 years (mean ± SD). Most were female (77.6%). The two groups differed in age, family location, education level, number of chronic diseases, sleep duration, joint VAS scores, swollen joint count, tender joint count, and ESR.

Figure 1A shows the QoL network of patients without depressive symptoms, and Fig. 1B shows the network structure of patients with depressive symptoms. Both network diagrams have 28 edges, of which 25 are non-zero edges, reflecting the high density of the network.

In both networks, PF and RP strongly associate with each other, as shown in Table S1. The NCT results showed no difference in the global strength of the two networks (S = 0.201, P = 0.673), but there were differences in their structure (M = 0.274, P = 0.029) and some specific edge strengths. Further analysis showed that the edges MH-PF, MH-RP, MH-BP, MH-GH, MH-VT, RP-VT, and GH-VT were more strongly associated in the network with depressive symptoms.

The nodes in the two networks with the highest EI values are different, as shown in Fig. 2. MH (1.100) has the highest EI value in the network with depressive symptoms, followed by SF (1.000) and VT (0.960). GH (1.200) has the highest EI value in the network without depressive symptoms, followed by MH (0.950) and VT (0.870). Figure S1 provides specific values.

Both network structures showed ideal accuracy and stability. The CS-C of the EI was 0.749 in the network without depressive symptoms and 0.673 in the network with depressive symptoms (see Fig. S2). The results of nonparametric bootstrapping showed that the 95% CIs of the edge weights were narrow and highly accurate (see Fig. S3).

As shown in Table 2, RA patients with depressive symptoms had lower scores on all QoL domains than those without depressive symptoms. In both networks, MH and VT have the highest predictability, meaning that the surrounding nodes can explain more than half of the variation in these two nodes.

Figure 3 shows the relationship between disease activity and QoL domains in RA patients. Disease activity is directly related to SF, VT, GH, BP, RP, and PF, with PF (edge weight = 0.237) having the strongest association with disease activity.

Discussion

First, we performed a network comparison of QoL between RA patients with and without depressive symptoms and found differences in network structure and some specific edge strengths. In exploring the relationship between disease activity and QoL domains, PF was found to have the strongest association with disease activity.

Our study found that MH, SF, and VT showed stronger associations in the network with depressive symptoms, and MH played an important role. Lempp et al. (2011) showed that both RA and depression can be detrimental to a patient’s mental health, but depression declines more rapidly in the mental health domain. When a patient’s mental health is impaired, it may affect their cognitive abilities and prevent them from interacting properly with others (Dobrek & Głowacka, 2023; Iaquinta & McCrone, 2015). According to the Attachment Model Theory, the inability to develop strong and lasting emotional connections with others can be one of the reasons for depressive symptoms (Bernaras, Jaureguizar, & Garaigordobil, 2019). As the depressive symptoms persist, the patient’s interest in things diminishes, which may affect their vitality (Lempp et al., 2011).

PF and RP showed strong associations in both networks. As the disease progresses, RA can lead to joint damage and even disability, which has a significant impact on a person’s health, especially in the domain of physical function (Lempp et al., 2011). Limitations in physical function can be associated with increased difficulty in completing daily activities and work (Citera et al., 2015). For example, some manual workers may not be able to perform heavy physical activities as normally. Even cerebral workers may be unable to perform fine hand activities, such as keyboarding. This may explain a strong association between PF and RP in patients with RA.

Significant negative association between disease activity and QoL, consistent with previous studies (Gouda et al., 2023; Qorolli et al., 2019).In addition, we found that PF was most strongly associated with disease activity. Disease activity is often accompanied by elevated levels of inflammation, which may produce a range of clinical symptoms, including joint pain, joint swelling, morning stiffness, and even systemic symptoms that limit physical function (Feng et al., 2024; Parrey, Koka & Ismail, 2024). It’s worth noting that patients with RA, influenced by limitations in physical function, may reduce activity, leading to a decrease in skeletal muscle content and an increase in fat content. Fat content has been associated with an increased inflammatory load, which can lead to increased levels of inflammation and disease activity in turn (Khoja et al., 2018; Sokka et al., 2008).

To our knowledge, this is the first study to explore the relationship between disease activity and QoL in RA patients using a network analysis approach. However, this study has some limitations. First, we recruited 424 RA patients from the First Affiliated Hospital of Anhui Medical University in Hefei, China; their small sample size does not allow for widespread generalization. Second, the cross-sectional study design didn’t permit causal inferences. This study can only provide potential hypotheses for future intervention studies and longitudinal studies. Third, we assessed QoL and depressive symptoms using self-rating scales; future studies may consider more objective assessment methods. Furthermore, the lack of the HADS anxiety subscale in this study limited its ability to assess comorbidity, which may have led to an underestimation of the psychological burden. Fourth, we did not collect treatment-related variables; treatments (e.g., traditional synthetic disease-modifying antirheumatic drugs/csDMARDs, biologics/bDMARDs, targeted synthetic disease-modifying antirheumatic drugs/tsDMARDs, glucocorticoids, etc.) may directly or indirectly affect patients’ QoL and depressive symptoms. Fifth, the study design does not include a control group for comparison, which may differentiate the effects of the intervention from other confounding variables. To overcome these limitations, future studies could consider controlled trials with extended follow-up periods to strengthen causal inferences and assess long-term effects.

Conclusion

Disease activity is strongly associated with QoL in RA patients, especially in the domain of PF, and the structure of the QoL network changes in the presence of depressive symptoms in RA patients.

Supplemental Information