Major Depressive Disorder (MDD) is a prevalent mood disorder characterized by high prevalence, recurrent episodes, low remission rates, and high suicide rates (1–3). It is the second most prevalent cause of disability globally, affecting over 3.32 million individuals (4). MDD adversely affects not only social functioning and quality of life, but also significantly increases the risk of suicidal behaviors, thereby imposing substantial burdens on affected individuals, their families, and communities (5, 6). Suicide is the most severe consequence of depression. Globally, over 720,000 people die by suicide annually (7). An estimated 90% of individuals who die by suicide suffer from one or more mood disorders, with MDD accounting for 59-87% of all reported suicides (8). Suicidal ideation (SI) often presents prior to an index suicide attempt or fatality (9). In China, among individuals with MDD, the prevalence of SI is 27.5% over the past month and 53.1% over their lifetime (10). The greater the severity and pervasive the SI, the higher the likelihood that it will lead to an attempt (11). Therefore, it is necessary to further explore the risk factors for SI among patients with MDD in order to better prevent suicidal behaviors.

Somatic symptoms are common among MDD patients and present with diverse manifestations. In Western countries, approximately 66%–93% of individuals with MDD experience somatic symptoms of varying severity (12, 13). Under the influence of collectivist values, Chinese people are more likely to report somatic symptoms (14, 15). In China, over 70% of patients exhibit moderate to severe somatic symptoms (16). Even though certain somatic symptoms are not covered by the Diagnostic Statistical Manual (DSM-5), they often co-occur with depression (17). These symptoms are associated with more severe and longer-lasting depression, increased disability, poorer clinical outcomes, elevated medical expenses, and worse quality of life (12, 18–21). These factors also serve as independent risk factors for SI in individuals with MDD (22).

MDD patients with SI exhibit more frequent and severe somatic symptoms compared to those without SI (23). A Turkish study has shown that individuals with SI have an average of 20.1 somatic complaints, while those without SI have an average of 10.6 somatic complaints (24). Previous research has demonstrated a significant association between somatic symptoms and a heightened risk of SI, with this risk seeming to exist independently of the presence of concurrent mental disorders (17, 18). However, there is a lack of longitudinal research evidence regarding individuals with MDD, as well as an exploration of whether there is a non-linear relationship between somatic symptoms and SI.

Somatic symptoms in patients with depression can be categorized into different subtypes, including pain symptoms, autonomic nervous system symptoms, energy- symptoms, and central nervous system (CNS) symptoms (20, 25). Most prior studies on the association between somatic symptom subtypes and SI have focused on single symptoms within specific somatic subtypes. However, patients often experience multiple somatic symptoms simultaneously rather than just one. Therefore, considering only single symptoms when examining the relationship between somatic symptoms and SI is insufficient. The cumulative impact and overall burden of co-occurring somatic symptoms must also be taken into account. While some studies have concentrated on the overall burden of pain symptoms, research on other subtypes remains limited.

Therefore, this longitudinal study of MDD patients in China, evaluating the burden of somatic symptoms and their subtypes (including pain, autonomic, energy, and CNS symptoms), aimed to: (1) investigate the associations between somatic symptoms and their subtypes with SI, and (2) evaluate non-linear relationships among somatic symptoms and their subtypes with SI.

In this longitudinal study, our findings suggest a significant association between somatic symptoms and their subtypes with SI among patients with MDD. Specifically, the somatic symptoms (e.g., total, painful, autonomic, and energy symptoms) significantly increased the risk of SI in a non-linear manner. In contrast, we observed linear associations between CNS symptoms and SI.

Our study found that total somatic symptoms were significantly associated with SI, independently of depressive symptom severity. These findings align with prior research (17, 23, 37–39). A meta-analysis of 33 studies demonstrated a significant association between somatic symptoms and increased risk of SI across diverse populations (37). Importantly, this increased risk was observed to be independent of any co-occurring mental disorders. Furthermore, a Japanese study focused on adolescents showed that increased somatic symptoms were significantly and independently associated with SI during mid-adolescence (38). This association remained significant after accounting for additional psychopathological and behavioral symptoms, as well as potential confounding factors. Another longitudinal study of 6,934 Chinese adolescents found that higher somatic symptom scores were more likely to report SI at baseline, with this tendency persisting at the 1-year follow-up (39). However, investigations focusing on patients with MDD remain confined to a handful of cross-sectional studies. For instance, a study of 217 Chinese patients with first-episode MDD found that those with more somatic symptoms had a greater risk of SI (17). Similarly, a Korean study of 811 outpatients with MDD found that patients with SI had significantly higher somatic symptom scores compared to those without SI (23). Although these studies varied in design, population, and assessment methods for somatic symptoms, SI, and risk factors, as well as in the number and type of somatic symptoms examined, their overall findings are consistent with ours. By using a longitudinal design and adjusting for multiple confounders—such as depression and anxiety severity, first-episode status, current and prior antidepressant use, sleep medications use, comorbid physical illness, and various health-related factors—our study provides more precise evidence on the association between somatic symptoms and SI in MDD patients.

Our study indicated that different subtypes of somatic symptoms (e.g., pain, autonomic, energy symptoms, and CNS symptoms) are significantly associated with SI. These results align with those of prior studies. For instance, a meta-analysis of 31 studies showed that individuals with any form of physical pain were more likely to report both current and lifetime SI, suicide plans, and attempts (9). A cross-sectional study conducted in Korea involving 414 patients with MDD found that patients with painful somatic symptoms had a 1.7 times higher risk of experiencing SI during the current depressive episode compared to those who did not report pain (40). Mao et al. studied MDD patients and found that specific pain symptoms (e.g., limb and pre-verbal pain) and certain autonomic symptoms (e.g., weight loss, increased appetite, hypersomnia, hyposexuality, and respiratory, circulatory, urinary, and sensory system symptoms) were significantly associated with SI (41). Jeon et al. found that MDD patients with SI exhibited significantly higher frequency and severity of somatic symptoms than those without SI, with current suicidal risk specifically associated with chest pain in men and neck or shoulder pain in women (23). Fang et al. found that among individuals with first-episode MDD, those with a higher number of somatic symptoms had an increased risk of SI, with pain symptoms, pre-verbal pain, and specific autonomic nervous system symptoms (e.g., late insomnia, weight loss, sensory system complaints, hypersomnia, and hyposexuality) associated with current SI (17). However, previous cross-sectional studies of MDD patients have mainly examined associations between single symptoms within specific somatic subtypes and SI, suggesting that a single-symptom perspective may underestimate the overall risk. Given that individuals with MDD frequently present with multiple concurrent somatic symptoms, the overall burden within each subtype requires systematic evaluation. This study offers a more comprehensive and accurate evidence by calculating total scores for four somatic subtypes to assess their overall burden and relationship with SI. This enables more effective identification of high - risk individuals and the development of more targeted management strategies.

In addition, we found a significant nonlinear relationship between total somatic symptoms and SI. When somatic symptoms are mild, the risk of SI remains low and stable. However, as symptoms worsen, the risk of SI increases significantly. This nonlinear relationship can help establish a treatment threshold. specifically, a score of 49 could serve as a practical cutoff for initiating somatic-targeted interventions in clinical practice. Similar nonlinear patterns were observed for pain, autonomic, and energy symptom subtypes. Defining distinct thresholds for each subtype enables more precise risk assessment. Notably, no nonlinear dose-response relationship was observed between CNS symptoms and SI. This may be related to the multidimensionality and complexity of central nervous system symptoms. The underlying mechanisms warrant further in-depth investigation.

Individuals with MDD who have somatic symptoms exhibit more complex clinical manifestations than those without somatic symptoms (42). These individuals usually have a more severe condition, a longer disease course, and a lower quality of life (16, 43). Persistent severe somatic symptoms not only indicate clinically significant impairment (44), but also significantly affect daily activities and functional levels (45). Furthermore, the severity of somatic symptoms may trigger health anxiety and concerns, which in turn may further increase the risk of SI (46).

The mechanistic association between somatic symptoms and SI in MDD is underpinned by multi-system dysregulation. Peripheral and central elevations of pro-inflammatory cytokines (TNF-α, IL-6) (47–52), hyperactivity of the HPA axis (47, 53), dysfunctions in the serotonergic and noradrenergic neurotransmitter systems (54–56), alterations in cerebral structure—such as regional grey-matter volume reductions and aberrant functional activity (47, 53, 57–59)—interact synergistically. These interactions form a complex pathophysiological network that influences the link between somatic symptoms and SI in patients with MDD.

Our results underscore the necessity of tackling somatic symptoms in MDD management. Early identification of clinical features that increase disease burden and management complexity is crucial for preventing depression from progressing to treatment-resistant and suicidal trajectories, as highlighted by Fiorillo et al. (60). Establishing subtype-specific thresholds allows for more precise risk stratification. In line with the international consensus by Maina et al. (61), patients with a high somatic burden should undergo multidimensional assessment and proactive suicide-risk monitoring. We recommend a total somatic score ≥49 as a practical threshold to trigger intensified surveillance and early initiation of symptom evaluation and targeted interventions, so as to prevent the progression of depression toward treatment resistance and suicidal trajectories. Clinicians should routinely and comprehensively assess somatic symptoms, incorporate symptom-targeted strategies into treatment plans, select antidepressants with analgesic benefits or minimal physical symptom exacerbation based on individual patient profiles (62), and combine these approaches with evidence-based non-pharmacological interventions, such as cognitive-behavioral therapy, exercise therapy, and other pharmacological treatments (63–65). Timely identification and effective management of somatic symptoms and their subtypes are essential for improving depression prognosis and reducing suicidal behavior.

Our study has several notable strengths. Firstly, to our knowledge, this is the first longitudinal study to explore the relationship between somatic symptoms and SI in patients with MDD. Secondly, our study examines the relationship between somatic symptoms and SI across four major subtypes: pain symptoms, autonomic nervous system symptoms, energy symptoms, and CNS symptoms. This is in contrast to previous studies that often focused on specific symptoms within subtypes. Thirdly, we identified potential inflection points in the non-linear relationships between somatic symptoms and SI. This is crucial for understanding the complex dynamics underlying these associations and informing targeted interventions.

When interpreting and generalizing our findings, several limitations must be considered. Firstly, the study was based in a hospital setting, which might restrict the results to be generalized to patients who do not seek treatment in hospitals. Secondly, we used the SSI scale to assess the 28 types of somatic symptoms in patients and did not measure or analyze other somatic symptoms that might be present in our study population. Thirdly, since we relied on self-report measures to assess somatic symptoms, SI, and depressive symptoms, recall bias could not be entirely avoided. Fourthly, although a longitudinal design was employed, the possibility of unmeasured confounders and reverse causation limits causal inferences.

Our study shows that somatic symptoms and their subtypes are significantly associated with SI in MDD patients. A critical threshold at ≥ 49 points on the total somatic symptom score indicates significantly elevated suicide risk. We recommend: (1) routine standardized assessment of somatic symptoms in all MDD patients;(2) implementation of targeted interventions for patients exceeding this threshold; and (3) enhanced suicide risk monitoring for these patients. This evidence-based stratified management approach holds significant clinical value for improving depression outcomes and suicide prevention.