Do Chronic Low Back Pain and Chronic Widespread Pain differ in their association with Depression Symptoms in the 1958 British Cohort?

Abstract Objective Depression frequently coexists with chronic pain. Contemporary models suggest that these conditions share pathobiological mechanisms, prompting a need to investigate their temporal association. This investigation aimed to explore two distinctly different chronic pain conditions, and their cross-sectional and prospective associations with depression. Methods Self-reported information was available on chronic widespread pain (CWP), chronic low back pain (CLBP) (45 years), and depression symptoms (45 and 50 years) from up to 9,377 participants in the 1958 British cohort. Depression symptom outcomes were derived by “Clinical Interview Schedule-Revised” (45 years) and “Short Form-36” (50 years). Relationships between both chronic pain conditions and depression symptoms were investigated by fitting four separate logistic regression models, each with varying levels of covariate adjustment, including depression at baseline. Results CWP was associated with depression symptoms cross-sectionally (odds ratio [OR] = 2.04, 95% confidence interval [CI] 1.65, 2.52; P < 0.001, n = 7,629), and prospectively when fully adjusted for baseline, sociodemographic, lifestyle, and health covariates (OR = 1.45, 95% CI 1.17, 1.80; P = < 0.001, n = 6,275). CLBP was associated with depression symptoms prospectively (full model: OR = 1.28, 95% CI 1.01, 1.61; P = 0.04, n = 6,288). In fully adjusted models the prospective association of CWP with depression symptoms was more heavily influenced by our covariates than CLBP with depression symptoms. Conclusion Pain may be a stressor from which depression can arise. Development of depression may be differentially dependant upon the type of pain experienced. Screening for depression symptoms among individuals with both chronic pain conditions is indicated and should be repeated over time.


Introduction
Chronic pain is prevalent and burdensome, with an estimated one third to half of UK adults experiencing chronic pain [1]. The economic burden of low back pain alone was calculated to be approximately £10 billion per annum in 2000 [2]. Similarly, depression is considered to be the largest contributor to global disability by the World Health Organization [3]. As of 2014 in the United Kingdom, 7.5% of the population was estimated to suffer from depression, with a peak in prevalence observed during middle adulthood [4]. Pain commonly coexists with psychiatric conditions, notably depression, often with devastating individual and societal consequences [5,6]. Critically, when presenting together depression and chronic pain yield poorer treatment responses and outcomes [7]. Recent research has focused on elucidating the mechanistic basis for the observed coexistence of these two conditions, prompting a need to explore their temporal relationships, particularly regarding potential differences between pain types (e.g., regional and global pain conditions). These relationships are of growing importance to public health policy given that chronic pain prevalence is predicted to rise due to the COVID-19 pandemic [8].
Previous studies have explored the associations between pain and depression cross-sectionally at a population level, and the potentially bidirectional nature of their relationship has been recognised [9]. Chronic pain is a physiological stressor and plausible antecedent to depression [10], yet to our knowledge there are no studies that explore and compare the link between different types of pain conditions and subsequent depression. Regional (e.g., chronic low back pain [CLBP]) and widespread pain (e.g., fibromyalgia) conditions can be considered distinct phenotypes of chronic pain, with observed differences in sensory, clinical, and psychosocial measures [11,12]. These differences are highlighted in the new IDC-11 classification codes, where chronic widespread pain (CWP) and chronic primary musculoskeletal pain (which includes chronic primary low back pain) are subgroups of chronic primary pain [13]. These subgroups have unique features that differentially overlap with the proposed mechanisms that link chronic pain and depression; therefore, it is important to further understand their independent relationships with depression.
Considerable research has examined the shared biological mechanisms of chronic pain and depression. Both the hypothalamic-pituitary-adrenal axis, and neuroimmune inflammatory mechanisms have been implicated as shared aetiology [14][15][16][17][18]. Recent research has demonstrated upregulated activation of glia in the brain of individuals with fibromyalgia (compared to healthy controls) suggesting that neuroimmune pathways are implicated in CWP conditions [19]. Furthermore, there is evidence of greater neuroimmune activation in the brain of persons with CLBP with radiculopathy (more akin to CWP) compared to those with axial CLBP (a regional pain condition) [20]. This neuroimmune activation was positively correlated with scores from the American College of Rheumatology Fibromyalgia Survey Criteria from each CLBP subgroup, suggesting it may be implicated in widespread symptom profiles, and differ from that of regional symptom profiles [20]. Evidence of differential relationships between pain conditions and depression, will not only provide further evidence for proposed pathobiological mechanisms linking the conditions, but crucially may inform public health policy and treatment pathways such as depression screening protocols in primary healthcare settings, and the provision of integrated health services.
The objectives of this study are to investigate the cross-sectional and prospective associations between CWP and CLBP with depression symptoms in mid-life, using the 45 year (baseline) and 50 year (follow-up) sweeps of the 1958 British Birth Cohort (1958BC, otherwise known as the National Child Development Study). We also aimed to explore how the associations are influenced by socio-demographic, lifestyle, and health factors. We hypothesised that both CWP and CLBP increase the risk of depression symptoms in mid-life. Given the greater overlap of proposed mechanisms between CWP and depression, we hypothesised that these variables would be more closely associated than CLBP and depression, both cross-sectionally and longitudinally.

Methods
We utilised the 1958BC as it is a rich longitudinal source of health, lifestyle, and sociodemographic data highly relevant to the aims of our investigation, including comprehensive pain and mental health data from participants aged 45 and 50 years (y) [21]. The 1958BC comprised 17,638 individuals born in one week in 1958 across England, Scotland, and Wales, with the subsequent addition of 920 immigrants also born within the same week [22]. Cohort members have been followed since birth with data collection at multiple time points, including a biomedical survey involving objective physical assessment at age 45y [21]. In this study, we used data from surveys at ages 45 and 50y, to construct our baseline (i.e., CLBP, CWP, and depressive symptoms aged 45y) and prospective (i.e., depressive symptoms aged 50y) measures. Data from surveys at ages 33, 42, and 45 were used to provide information on covariates. From the cohort that completed the 45y biomedical survey (n ¼ 9377), the current cross-sectional analyses were restricted to 7629 individuals with information for CWP and depression symptoms (45y), and 7643 individuals with data on CLBP and depression symptoms (45y) (Supplementary Data Figure 1). Prospectively, data were available for 6275 individuals with CWP and depression symptoms (50y); and 6288 individuals with data on CLBP and depression symptoms (50y).

Depression Symptom Measures
The depression symptom variable (45y) was constructed using a summary score based on the "Clinical Interview Schedule-Revised" (CIS-R) which is a screening instrument for depression, and designed for use by lay persons. In the 1958BC this assessment was conducted using computer assisted personal interview (CAPI) by research nurses [21]. The CIS-R has been previously validated, and comprises multiple sections pertaining to psychiatric symptoms evident within the previous month, which are allocated a numeric score (depression summary score ranging from 0 to 4, based on severity and frequency of symptoms in the preceding week) [23][24][25]. We constructed a categorical yes/no indicator of depression symptoms (aged 45y) using a cutoff score of 2 for the CIS-R, consistent with previous research [23]. From the 1958BC at age 50y data regarding depressive symptoms was available from an alternate assessment tool, the 36item "Short-Form Health Survey" (SF-36). The SF-36 is a self-completed, paper-based measurement tool yielding information on three primary dimensions: functional status, well-being, and overall evaluation of health, and embodies the Mental Health Index (MHI-5) [26]. The MHI-5 has been validated previously as a means of screening for symptoms of depression and provides scores ranging from 0 to 100 [27]. Again, a categorical yes/no indicator of depression symptoms was constructed using cut-off scores consistent with prior research (52 indicating depressive symptoms) [28].

Chronic Widespread Pain and Chronic Low Back Pain Measures
CWP and CLBP were each obtained from the biomedical survey via a construct of self-reported pain related information [21]. Questions pertained to the duration of painful symptoms (delineating acute vs. chronic pain, defined here as pain lasting longer than three months), and a body chart on which the cohort member was required to shade in the painful regions [29]. From this, multiple pain location variables were derived enabling construction of the CWP variable, such that it aligned with the following subjective pain related component of fibromyalgia 1990 diagnostic criteria: "pain lasting more than three months that is bilateral, axial, and present above and below the waist" [30]. Specifically, our CWP measure included individuals with painful symptoms in their lower body (i.e., below the gluteal folds), trunk and/or upper limb regions. The CLBP variable was constructed in keeping with the following definition: "Low back pain is pain, muscle tension, or stiffness localized below the costal margin and above the inferior gluteal folds" and was considered chronic when persisting for longer than 12 weeks [31]. Symptoms extending into the legs were excluded from our CLBP measure, as were individuals meeting our aforementioned CWP criteria. In doing so, our approach ensured that the two pain cohorts were distinct from each other, with no overlapping cases.

Sociodemographic, Lifestyle, and Health Covariates
The rationale for the selection of covariates was based on existing evidence of association with the respective exposures (CWP and CLBP) and the outcome (depression symptoms) [32]. Covariate measures were derived from self-completed questionnaires aged 45y (i.e., general health, screen time for leisure, social contact, active social group membership, bill payment difficulty, and separation status). Computer assisted self-interview (i.e., for alcohol intake aged 45y), computer assisted personal interview (i.e., for smoking status and physical activity aged 45y, social class aged 42y), and nurse-led physical testing (i.e., weight and height aged 45y) were also sources of covariate information. Social class was based on the Registrar General's classification: professional and managerial/technical, skilled non-manual and manual, and unskilled and other/unknown [33]. The highest qualification was constructed using maximum qualifications aged 33y (derived variable from interview) and complemented with information on lifelong learning from the 42y survey [21]. This categorization included nil qualifications, secondary, and tertiary, and was consistent with prior research. Separation status was categorised into yes/no, and bill payment difficulty into minor/no difficulty, difficulty, and great difficulty. Social contact (regularity of contact with friends or acquaintances) was categorised as multiple times per month, better reflecting data distribution and reducing modelling parameters [34]. Active social group membership (i.e., a member of a social/recreational group, trade unions, commercial/professional organisations, political/advocacy/cultural or sporting group) was categorised as yes/no. Screen-time for leisure was constructed combining PC and TV utilization, and categorised as non-sedentary ( 4 hours) and sedentary (> 4 hours). Alcohol intake was categorised as never/rarely, light/moderate, and heavy consumption, while general health was either excellent, good, or fair. Body Mass Index (BMI) was calculated as weight (kg)/ height (m) 2 aged 45y. BMI categorisation was based on the World Health Organization classification [35], and included underweight (<18.5), normal (18.5 to <25), overweight (25 to <30), and obese (30). Physical activity was constructed using three related measures, regular exercise participation, exercise frequency, and physical demands of work. Categories include four times per month or more and three times per month or less, to better reflect data distribution. Smoking categories include current, ex/occasional, and never.

Statistical Analyses
We explored descriptive distributions by tabulating frequencies and v 2 testing to compare bivariate relationships. Relationships of the chronic pain conditions with depression symptoms were investigated by fitting four separate logistic regression models, each with a different level of covariate adjustment guided by a backward stepwise modelling approach [36]. For cross-sectional analyses of CWP and CLBP associations with depression symptoms (45y), the four levels of adjustment were as follows: (1) Baseline: sex; (2) Socio-demographic: sex, social class, highest education level, separation status, bill payment difficulty; (3) Lifestyle: sex, smoking status, physical activity, alcohol intake, social contact, active member of a social group; (4) Health: sex, and general health. Similar covariate adjustment was used for prospective associations of CWP and CLBP with depression symptoms (50 years), with physical activity being replaced by sedentary status and all models additionally adjusted for baseline depression symptoms and BMI. Complete case analysis was used throughout our analyses. As sensitivity analyses, we imputed missing covariate data using using 20 imputed datasets, and re-iterated all the analyses [37]. All analyses were conducted in STATA 14 (StataCorp LP, Texas USA).

Results
Our sample included 8,522 individuals with data available for CWP and depressive symptoms, and 8,539 individuals with data available for CLBP and depressive symptoms, both at age 45y. Of the first group, 1,050 (12.3%) of individuals reported CWP at age 45y. Depression rates in this group were stable over time, with 18.5% of individuals having depressive symptoms at 45y, and 18.6% at 50 years. For the CLBP and depression group, 1020 (11.9%) reported CLBP at 45y. Of this group, 7.9% experienced depression at 45y, which increased to 15% at age 50y. Supplementary Data Figure  1 depicts the participant flow through the study. Our descriptive analyses further demonstrated that the prevalence of CWP was associated with lower social class, lower education level, higher bill payment difficulty, smoking, reduced physical activity, increased screen time, lower alcohol intake, not being an active social group member, poor general health, and higher BMI (Supplementary Data Table 1). CLBP was associated with lower education level, smoking, and poor general health (Table 1, Supplementary Data Table 1). Depressive symptoms at ages 45 and 50y were more prevalent in women than in men (Table 1,  Supplementary Data Table 1). Lower education level, lower social class, higher bill payment difficulty, smoking, reduced physical activity, reduced social contact, not being an active social group member, increased screen time, lower alcohol intake, and poor general health were all associated with depressive symptoms aged 45 and 50y ( CWP was also associated with depression symptoms (50 years) in prospective analyses, which controlled for depression symptoms at baseline (OR ¼ 2.17, 95% CI 1.78, 2.65; n ¼ 6,275) ( Table 2). Adjustment for sociodemographic, lifestyle, and health related covariates each attenuated, but did not fully explain, the prospective association between CWP and depression symptoms (OR ¼ 1.45, 95% CI 1.17, 1.80; n ¼ 6,275). Fully adjusted sensitivity analysis using multiple imputation provided very similar results (OR ¼ 1.44, 95% CI 1.18, 1.77; n ¼ 6,943).

Discussion
To our knowledge this is the first study investigating the association between chronic pain and depression symptoms in mid-life, within the 1958BC. We found that CWP was strongly associated with depression symptoms both cross-sectionally and prospectively. This association was maintained when controlling for socio-demographic, lifestyle, and health covariates. In contrast, CLBP was only prospectively associated with depression symptoms. While both chronic pain conditions were prospectively associated with depression symptoms in the fully adjusted models, CWP was more heavily influenced by covariate adjustments than CLBP. That CWP was associated both cross-sectionally and prospectively with depression symptoms suggests they may share common pathobiological mechanisms.
The cross-sectional association between CWP and depression symptoms is consistent with previous research [38]. In our cohort, people with CWP at age 45 had a 45% higher risk of developing depression symptoms by 50y. This is lower than a previous study, which found that people with fibromyalgia (n ¼ 25,969) are over seven times more likely to be diagnosed with depression in the following 3-11 years [39]. The lower risk in the current study may be due to differences between CWP and fibromyalgia. Diagnosis of fibromyalgia requires not only widespread pain, but also additional somatic symptoms (e.g., fatigue, abdominal symptoms) [40]. Moreover, previous research has demonstrated higher levels of psychological distress in persons with meeting the diagnostic criteria for fibromyalgia, compared to those with CWP alone [41]. Finally, the shorter follow-up period of this study may influence effect sizes. The present study found no cross-sectional association between CLBP and depression aged 45y. This is in contrast with a cross-sectional Canadian study (n ¼ 118,000) in which people with CLBP were 6.2 times more likely to experience depression than those without pain [42]. Further, in the United Kingdom Keeley et al. (2008) found 25% of CLBP (n ¼ 108) patients suffered from comorbid depressive symptoms [43]. Interestingly, both studies used a time frame of 6 months for pain chronicity (compared to 3 months in this study). The differences in inclusion criteria may explain the inconsistent results, given that that changes in pain levels over a 3-month period can predict changes in depression outcomes [44]. Nevertheless, our prospective analysis implicates CLBP as a predictor of subsequent depression symptoms, which is consistent with previous research efforts [45]. Additionally, we were able to control for a number of sociodemographic, lifestyle, and health covariates in our analysis; these factors attenuated, but did not fully explain, the relationships between the two pain conditions and depression symptoms.
Multiple models of the chronic pain and depression relationship exist in the literature, many indicating that the two conditions have shared pathobiological mechanisms [10]. Considerable evidence from prior research supports dysregulation of the hypothalamic-pituitaryadrenal axis as a plausible biological pathway linking chronic pain and depression [18]. Similarly, recent research suggests that both pain and psychological stress can initiate a cascade of cellular reactions whereby neuro-immune (i.e., microglial) mechanisms are upregulated, leading to depression [46]. Our results support a model in which chronic pain is a stressor from which depression can develop, which culminates differently depending on the nature of the pain experience (i.e., regional compared to systemic). Stress response systems seem a likely pathway moderating the relationship of these conditions.
The results of this study have significant clinical implications. The longitudinal relationship between chronic pain and depression indicates that repeated screening for depression is required. This is particularly important for regional pain conditions (such as CLBP), as depression may develop over time; if screening only occurs at diagnosis, subsequent development of depression could be missed. Additionally, the strong association between pain conditions and depression, as well as that people experiencing both pain and depression have an attenuated response to treatment, indicates that this group may merit special consideration with regards to treatment pathways-early referral to multi-disciplinary care may be appropriate for this population. Socio-demographic, lifestyle and health covariates had a considerable influence on the prospective relationship between CWP and depression. This highlights the importance of such factors in influencing the health outcomes in people with CWP.

Limitations
As with all observational studies, despite comprehensive covariate adjustment we are unable to discount effects by residual confounding, and therefore associations observed in our study may not be causal. Secondly, while in line with previously validated methodologies [47,48], there are also limitations with the construction of both chronic pain measures. Namely, we did not integrate pain intensity, higher levels of which have previously been associated with greater depression symptoms [49], which may have affected our findings. Furthermore, chronic pain has known associations with other psychiatric conditions (e.g., post-traumatic stress disorder) which may confound our findings. Presence of multiple psychiatric conditions may be a confounder [50]. Our CLBP measure excluded individuals with symptoms extending to their legs. This was done to ensure that the CLBP and CWP conditions were phenotypically distinct, and to exclude potential cases of lumbo-genic radiculopathy, as neuropathic pain typically has a greater association with adverse mental health than somatic chronic pain alone [51]. We sought to mitigate the influence of neuropathic pain as a potential confounder for CLBP, though we were unable to do this for CWP. This may contribute to the weaker associations between CLBP and depression symptoms. Also, the study design did not account for comorbid musculo-skeletal complaints such as neck pain or other comorbid multi-joint pain, which commonly coexists with both low back pain and depression [52,53]. Regarding CWP, our measure closely aligns with the diagnostic criteria for fibromyalgia but does not include other symptoms required for diagnosis (e.g., measures of fatigue, cognitive issues). Caution is needed when aligning the CWP results with fibromyalgia, as the CWP variable constructed here may also be comparable to other conditions where widespread pain is a feature, for example auto-immune polyarthropathies [54]. While our results support a temporal relationship of chronic pain and downstream development of depression, our study does not account for the possibility of depression preceding the onset of pain and predisposing individuals to subsequent depression [10]. Finally, the depression symptoms variable at the two timepoints were constructed using different measures. However, both measures have been validated. The CIS-R has been validated against an interviewer-led version of the same for depression [55], while the MHI-5 has been validated for detecting mood disorders and psychological distress in various populations [28,56].  In conclusion, both CWP and CLBP appear to have a temporal and differential relationship with the latter onset of depression, where CWP is associated with depression symptoms cross-sectionally and prospectively; and CLBP is only associated with depression symptoms prospectively. Concomitance between CWP and affective mechanisms is suggested. Health, lifestyle, and sociodemographic factors were observed to influence the associations of both pain conditions with depression. Our results have implications for clinical management of individuals with chronic pain. Considerable attention to a chronic pain sufferer's psychological wellbeing is needed, and formalised screening for depression symptoms among this demographic may be indicated for best practice. Given the temporal relationship, this screening should be periodically repeated over time. Moreover, given that the chronic pain and depression frequently cooccur, dedicated multi-disciplinary treatment pathways should be developed to allow co-treatment of both conditions.

Authors' Contributions
All listed authors have made substantial contributions to the conception and design of the study, and/or data acquisition, analysis and interpretation of data, drafting of the manuscript, and final approval of the submission.

Research Ethics Approval
This work comprises secondary analyses of an anonymized data set and as such, ethical approval was not required. Original ethical approval for the 1958BC for the 45-year and 50-year data sweeps under study was granted by the South East Multi-Centre Research Ethics Committee (reference 01/1/44), and the London Multi-Centre Research Ethics Committee (reference 08/H0718/29), respectively.