Discussion

The main findings of this study are as follows. First, anxiety severity demonstrated an independent positive association with risk of RSA in patients with MDD; compared with patients with MDD and mild anxiety, those with significant anxiety and severe anxiety had 2.9 times and 11.8 times higher risks of RSA, respectively. Second, among patients with MDD and a history of RSA, higher levels of depressive symptoms, positive psychosis symptoms and thyroid function indicators (FT3 and TSH) are independently associated with increased anxiety severity.

This study assessed anxiety symptoms in 1718 patients with MDD, using the HRSA scale, with 1717 patients evaluated as having varying degrees of anxiety, indicating an extremely high comorbidity rate of anxiety in the MDD population. However, this finding raises an important question: whether the HRSA scale may be overly sensitive in assessing anxiety, potentially leading to an overestimation of comorbid anxiety in patients with depression. The result profoundly highlights the methodological limitations of relying on a single symptom scale in MDD: the high sensitivity of the HRSA primarily stems from its inherent overlap with core MDD symptoms (such as psychic anxiety, insomnia and somatic discomfort), which compromises its specificity. Consequently, what is being measured is essentially the anxiety symptom dimension within MDD, rather than a truly independent comorbid condition. The resulting ceiling effect homogenises the sample in terms of the anxiety variable, substantially weakening the scale’s discriminant validity and the precision of treatment outcome evaluation. Moreover, the pervasiveness of anxiety symptoms suggests that they may be an intrinsic component of the MDD disease phenotype, thereby diluting the clinical significance of stratifying patients or conducting association studies based solely on HRSA scores. Therefore, this phenomenon underscores the necessity, in complex psychopathology research, of employing a multi-faceted approach – integrating clinical interviews, multidimensional scales and factor analysis – to more accurately delineate the clinical meaning of symptom structures.

Research on the correlation between anxiety and suicide risk has yielded mixed results. For example, one study indicated that individuals with anxiety disorders were five times more likely to attempt suicide compared with those without. ^{Reference Moitra, Santomauro, Degenhardt, Collins, Whiteford and Vos16} In contrast, other studies have found no significant association between anxiety and suicidal behaviour, and some have even reported the opposite, suggesting that anxiety may be linked to a lower risk of suicide. ^{Reference Nock, Hwang, Sampson and Kessler17} The discrepancies across these findings may be attributed to factors such as methodological variations, differences in inclusion/exclusion criteria and sample heterogeneity. Furthermore, in some previous studies, a significant temporal gap between the assessment of anxiety levels and the occurrence of suicidal behaviour may weaken the observed correlation between the current state of anxiety and suicide risk. It is also important to note that although both suicidal ideation and suicidal behaviour are categorised under suicide risk, they represent distinct constructs. Core principles of the ideation-to-action framework suggest that the transition from having no suicidal ideation to developing ideation, and from ideation to engaging in suicidal behaviour, are distinct processes with different underlying mechanisms. ^{Reference Klonsky, Dixon-Luinenburg and May18} In this study, the severity of anxiety in patients with MDD and a history of suicide attempts within the past month was assessed with the HRSA. We found that the severity of anxiety symptoms was significantly associated with suicide risk in patients with MDD. Specifically, patients with MDD and severe anxiety exhibited a more than ten-fold higher suicide risk compared with those with mild anxiety. The mechanisms underlying the influence of anxiety on suicide are complex. Anxiety sensitivity may play a role in mediating the impact of anxiety on suicidal behaviour. ^{Reference Stanley, Boffa, Rogers, Hom, Albanese and Chu19} Research has confirmed an association between anxiety sensitivity and suicide, with significantly elevated anxiety sensitivity levels observed in individuals who have attempted suicide. ^{Reference Demirkol, Tamam, Namli, Karaytuğ and Yeşiloğlu20} Anxiety sensitivity may exacerbate the experience of anxiety and, in patients at high risk of suicide, could facilitate the progression from suicidal ideation to suicidal behaviour. ^{Reference Allan, Gorka, Saulnier and Bryan21} Furthermore, it may also influence suicide risk through indirect pathways. For instance, its role in promoting maladaptive coping strategies can increase the likelihood of developing suicidal ideation. ^{Reference Poorolajal, Haghtalab, Farhadi and Darvishi22–Reference DeMartini and Carey24}

Although this study cannot establish a causal relationship between depression severity and anxiety severity, their significant positive correlation has been consistently demonstrated here and in numerous other studies, highlighting the need for increased clinical attention to this comorbidity. This association may stem from shared neurobiological underpinnings interacting with environmental stressors, leading the two disorders to frequently influence and exacerbate each other. ^{Reference Coussement, De Longueville and Heeren25,Reference Zhu, Yin, Ma and Zhang26} Furthermore, our findings revealed a positive correlation between positive psychotic symptoms and anxiety severity, independent of depressive symptoms. This aligns with observations by Temmingh et al in schizophrenia populations. ^{Reference Temmingh and Stein27} The mechanisms underlying this link may involve multiple pathways, with sleep disturbances potentially serving as a mediating factor. Evidence suggests that sleep impairment is closely associated with the worsening of positive symptoms and can significantly intensify anxiety. ^{Reference Blanchard, Andrea, Orth, Savage and Bennett28} Therefore, future studies should include sleep quality as a key factor in both clinical evaluation and intervention strategies, to better understand its specific contribution to the link between positive symptoms and anxiety severity.

Additionally, the association between TSH levels and anxiety severity observed in the present study is consistent with earlier findings by Zhu et al in patients with MDD, further supporting the hypothesis that thyroid axis function may be linked to mood disorders. ^{Reference Zhu, Yin, Ma and Zhang26} From a neurobiological perspective, it has been proposed that thyroid hormones might influence the plasticity of key emotional regulatory brain regions such as the hippocampus, thereby potentially contributing to the pathophysiology of anxiety. Specifically, thyroid hormones are known to modulate the expression of neurotrophic factors, including brain-derived neurotrophic factor (BDNF), which plays a critical role in neuronal survival, differentiation and synaptic plasticity. ^{Reference Yu, Zhou, Yang, Jiang, Wang and Lv29} Thus, thyroid dysfunction may indirectly affect the functional integration and adaptive changes within emotional neural networks by regulating BDNF-related signalling pathways, possibly exacerbating or sustaining anxiety symptoms. It is noteworthy that the aforementioned mechanistic explanation remains speculative and has not yet been directly validated in populations with anxiety disorders. Current understanding is largely derived from animal models and neurophysiological studies in individuals with thyroid diseases, and the specific pathways and modes of action in human anxiety disorders require further investigation through longitudinal and mechanistic research. Future studies should integrate multi-modal neuroimaging, dynamic monitoring of peripheral biomarkers and genetic tools to more systematically explore the role of thyroid–brain interactions in the development and progression of anxiety disorders. Such approaches could provide a more solid foundation for understanding disease mechanisms and potential intervention strategies.

Our findings support a close association between anxiety and depressive symptoms, suggesting that in the clinical management of MDD, the assessment of anxiety severity should not be viewed merely as a descriptor of comorbidity, but should be integrated as a key dimension in core risk stratification and treatment decision-making. Elevated anxiety levels may identify a more complex and vulnerable MDD subgroup, characterised by more severe depressive symptoms, greater functional impairment, more prominent sleep and somatic complaints, and a potentially poorer initial treatment response. Therefore, systematic screening using conventional depression assessments alongside targeted anxiety scales, such as the Generalised Anxiety Disorder scale, can facilitate more precise disease stratification and prognosis estimation. In terms of management, identifying and actively addressing significant anxiety is crucial for optimising overall treatment strategies. Evidence indicates that patients with MDD and severe anxiety may require more nuanced therapeutic approaches; for instance, selecting antidepressants with efficacy across anxiety spectrum symptoms while considering potential activating side-effects, and integrating anxiety-focused cognitive–behavioural or mindfulness-based techniques within psychotherapy to address emotional distress more comprehensively. Furthermore, the observed links among anxiety, positive psychotic symptoms and neuroendocrine indicators such as TSH underscore the necessity of a transdiagnostic, multi-system perspective. Clinicians managing patients with MDD and prominent anxiety should remain alert to broader psychopathological network activity, including the assessment of psychotic symptoms and consideration of physiological tests such as thyroid function. In summary, dynamically incorporating anxiety severity into the holistic assessment and management framework for MDD not only promotes more individualised and proactive care but also provides a clinical foundation for shifting from a symptom-specific treatment model toward dimension-based transdiagnostic intervention. Future studies should further clarify the long-term outcomes of the high-anxiety MDD subtype and validate the effectiveness of integrated intervention strategies in this population.

This study has several limitations that should be noted. First, the cross-sectional design restricts causal interpretations regarding anxiety and RSA, as well as the relationships between other clinical factors and anxiety. Second, the demographic characteristics and the observed prevalence of anxiety in this study may be influenced by cultural or sampling factors. For instance, nearly all participants in this study exhibited mild anxiety symptoms, and very few were without anxiety, which may limit the generalisability of conclusions regarding the extent of anxiety’s role in RSA. Finally, suicide risk was assessed through face-to-face interviews in this study; the use of structured suicide risk assessment scales could help reduce potential bias in such evaluations.

In summary, our findings highlight an independent positive association between anxiety severity and RSA in patients with MDD, emphasising the clinical importance of managing anxiety symptoms in this population. This study also reveals independent correlations of anxiety with depressive symptoms, positive psychotic symptoms and thyroid hormone levels. Future research should systematically examine the differential effects of distinct dimensions of anxiety symptoms, such as psychic anxiety and somatic anxiety, as well as specific anxiety-related behaviours, including avoidance and hypervigilance, on the risk of suicide attempts in MDD. Such investigations will not only help clarify the specific pathways through which anxiety contributes to suicidal behaviour in MDD, but also provide an empirical basis for developing precise early-warning indicators and targeted symptom-specific interventions for high-risk individuals, thereby supporting the implementation of more stratified and personalised comprehensive treatment strategies in clinical practice.