Our objective was to expand current understanding on the epidemiology of depression and/or anxiety among patients with AS by synthesizing evidence on the risk and determinants of these psychiatric conditions. Altogether, a pooled relative risk of 1.51 (95% CI 1.28 to 1.79), suggesting a 51% higher risk of depression among individuals with AS compared to those without AS, highlights the importance of monitoring and identifying at risk patients during encounters of care. Our systematic review also identified areas where further work is needed. For example, only one included study assessed the risk of anxiety among AS patients (20), thus calling for future confirmatory studies. Limited research on determinants of depression and/or anxiety risk among AS patients is also a gap as these represent potential ways for identifying at patients who may be at risk for these conditions.
To our knowledge, this is the first systematic review on the incidence of depression and/or anxiety among patients with AS, which builds on prior systematic reviews on the prevalence of depression in this patient population (14). In their 2018 systematic review, Zhao et al. (14) identified 16 studies and reported the prevalence of depression ranging from 11% to 64% and pooled prevalence of 0.29 (95% CI 0.15 to 0.44), which is consistent with secondary outcomes in our current systematic review. The issue of prevalent (or comorbid) depression is particularly important in AS management, given associations with disease activity and functional impairment (14) as well as the diagnosis delay from the symptom onset (26). Equally important is the onset of depression following AS diagnosis and as we demonstrated in our systematic review, there is a 51% higher risk of depression among AS patients compared to individuals without AS. In addition, the only study assessing the risk of anxiety in patients with AS by Shen et al. (20) reported an 85% increased risk for anxiety, after adjusting for confounding, in comparison to controls. A number of biologic mechanisms may explain this increased risk and evidence suggests that the inflammatory nature of AS could be closely tied to depression and anxiety. As observed in patients with AS (27), individuals with depression and anxiety have also been shown to have elevated levels of proinflammatory cytokines, including tumour necrosis factor alpha and interleukin-6 (28, 29).
Sex-specific risk estimates for depression and anxiety reported (19, 22) or available upon request (20) from three studies allowed us to examine potential differences between men and women. This may be particularly relevant given the predominance of AS among men than in women (4-6) that contrasts the predominance of psychiatric conditions among women (30, 31). Two of the studies included our review provided unadjusted risk estimates for depression with conflicting results, with one suggesting no increased risk when stratified by sex (19), and the other reporting significantly increased risk for depression among both men and women with AS (22). Although both of these studies used administrative health data, they differed in the ICD codes used to define depression. Indeed, the study by Meesters et al. (22) that noted significant risk estimates for men and women used a broader set of ICD codes to define depression. In contrast to Sundquist et al.’s (19) insignificant sex-specific risk estimates for severe depression in patients with AS, when the same authors calculated risk estimates for affective disorders (including depression, dysthymia, and depression not elsewhere classified), standardized incidence ratios for men (1.64, 95% CI 1.21 to 2.16) and women (1.82, 95% CI 1.20 to 2.65) reached significance. The third article reported adjusted sex-specific risk estimates for depression and anxiety notably showed that men with AS have more than a two-fold increased risk for incident anxiety whereas women with AS have no increased risk, and further reported that the risk of depression was significant after stratification, but higher among males (20). Taken together, these findings emphasize the need for future research to clarify the effect of sex on the risk of mental disorders in AS, and highlight the potential for tailored approaches that address gender differences when managing mental health in patients with AS.
We additionally synthesized evidence on determinants of depression and/or anxiety among patients with AS, as these may represent potential targets for intervention or means for identifying high-risk patients. Only two studies were included, based on our a priori definition of a determinant as a factor having independent association with depression and/or anxiety based on multivariable regression methods. Aside from this limited number of studies, it is also important to note that both included studies used cross-sectional designs and as such, temporality cannot be determined. Nonetheless, associations between measures of AS disease severity and depression, and quality of life and both depression and anxiety (24) point to potential impacts on mental health of managing AS. As well, a 10-fold higher odds of depression associated with smoking reported in one study (23) may suggest a modifiable target. Finally, although two of the studies assessing incidence of depression and anxiety indicated factors they adjusted for such as comorbidities, urbanization, income (20), and outpatient care utilization, obesity conditions, smoking (21) – risk estimates for these associations were not reported, precluding the ability to assess whether any of these were potential determinants of depression and/or anxiety.
Strengths and limitations of our systematic review deserve discussion. We collaborated with an information scientist to develop the database search strategies, with the information scientist executing all of the searches. However, the inclusion of relevant studies may have been limited by publication bias as in any other systematic review. We did not consider abstracts given the importance of being able to assess the quality of the included studies. We also did not consider grey literature, although this may not be problematic given the specificity of our topic, that is the incidence of depression and/or anxiety in AS, lends itself to peer-reviewed scientific manuscripts.
Given the demonstrated association between psychiatric complications, namely depression, and disease activity in AS (14), clinical implications of our systematic review also warrant discussion. In particular, our findings suggesting a 51% higher risk of depression among individuals with AS compared to those without AS, highlights the importance of identifying at risk patients during encounters of care. Though our synthesis indicates that evidence on predictors of depression in AS are limited, there is data suggesting AS-related factors (e.g. disease activity) as potential markers for depression risk (24). Another consideration is the treatment of psychiatric complications in AS, which to our knowledge, has also received very little attention. In their 2016 study on the prevalence of depression and anxiety among patients with rheumatic diseases (n = 514) including 44 with AS, Anyfanti et al. also assessed treatment of psychiatric complications and reported considerable under-treatment with less than 10% of AS patients receiving an antidepressant or antianxiolytic (32). Finally, given the potential role of inflammation in the relationship between AS and depression (27-29), treatments for AS, particularly anti-tumour necrosis factor (anti-TNF) agents, may also impact depression as suggested in prior clinical study (33).