Based on the data analysis of Employees ' Health Cohort Study of Iran (EHCSIR), people with lifetime MDD or past twelve months MDD, have a significantly higher risk of susceptibility to COVID-19 infection compared to non-sufferers. Among confirmed COVID-19 cases, we found no evidence of increased risk of hospitalization in participants with lifetime MDD or past twelve months MDD.
There is controversy on the increased risk of susceptibility to COVID-19 infection among MDD cases. In one meta-analysis on the association of preexisting mental or neurological disorders with COVID-19 susceptibility, 18 studies on preexisting mental disorders were recruited. Six studies of them with more than 62 million participants showed that patients with mood disorders have higher risk of susceptibility to COVID-19 infection (OR = 2.02; 95% CI: 1.08–3.76) [22]. In another meta-analysis on 21 studies with more than 91 million individuals, no association was found between mood disorders and COVID-19 susceptibility (OR = 1.27; 95% CI: 0.73–2.19; n = 65,514,469) [3]. Only two of the 18 recruited studies in the first meta-analysis and one study out of 21 studies in the second meta-analysis were from low-middle countries [3, 22].
The difference in the findings of the studies can be due to the difference in demographic, socio-economic and clinical characteristics. In addition, since the biological and neural mechanisms of different mental disorders may be different, it can lead to confounding association [22].
Several causal pathways have been proposed to explain the increased risk of susceptibility to COVID-19 infection in people with depressive disorders. The first group of causal pathways are MDD-related behaviors, which include inhibition, apathy, lack of motivation, cognitive deficits [3], reduced awareness of risks [23], neglect of self-protection [4], less confidence in their decisions and government messages to stay at home [24], sleep dysregulation, habitual inactivity [3], impairment in evaluating health information and obtaining preventive behaviors [25], stigma and accessing health care [26]. The second group is comorbidities associated with MDD, which includes drug use disorders [27], diabetes [28], obesity and cardiovascular diseases [3]. The Third group is biological factors, which include inflammation [14], activation of the hypothalamic-pituitary-adrenal axis [29], changes in corticosteroids [30], drug therapy [3], and shared genetic susceptibility factors [31, 32]. On the other hand, social isolation, unemployment, and low interpersonal contact [3, 6] are possible mediators that may explain lower risk of susceptibility to COVID-19 in MDD patients [6, 33].
Most of the previous studies showed an association between depressive disorders and COVID-19 hospitalization. A meta-analysis of 8 studies involving more than 25 million participants showed that preexisting mood disorders were associated with increased risk of illness severity (OR = 1.34; 95% CI: 1.08–1.67) [22]. Another systematic review and meta-analysis reported that COVID-19 hospitalization is significantly higher in those with mood disorders (OR = 1.31; 95% CI; 1.12–1.53) [3].
Lack of motivation and problems in evaluating health information in MDD patients can lead to non-adherence to preventive behaviors, including tertiary prevention [4]. Also, the low economic status in this group leads to delays in receiving medical care and going to the hospital [7]. Social determinants of health such as economic insecurity, health literacy, and limited access to healthcare have been describes as mediators of this association [3]. The negative finding in our study might be explained by the characteristics of our study population. Since all of the subjects of our cohort study were public employees and majority of them were working at hospitals or health centers, it is reasonable that their health literacy, economic stability, and access to healthcare services are higher than general population and are not affected by their mental disorder.
There are few comparable data on the relationship between psychiatric disorders and the risk of susceptibility to COVID-19 infection and hospitalization [34]. Research in the field of other respiratory viruses shows that mental disorders can affect the vulnerability to the COVID-19 [35]. Several experimental studies have reported a dose-response relationship between psychological stress and acute respiratory infectious disease [34]. Based on the analysis of the UK biobank cohort, the risk of susceptibility to COVID-19 infection increases with the increase in the number of mental illnesses, but there is no significant relationship with the time that has passed since the diagnosis of a mental disorder [34]. In a similar study in the US, patients with recently diagnosed (within the past year), are at a high risk of susceptibility to COVID-19 infection (AOR = 10.43), but the strength of this relationship is less for patients who have been diagnosed for more than a year (AOR = 1.48) [4].
Strengths and limitations
This study has several strengths. First, the most important strength of this study is that MDD was measured with a valid diagnostic tool, not based on hospital records or self-report or symptom scales. In all of the similar published studies that we found, the pre-existing major depressive disorder was assessed based on diagnosed, treated, or hospitalized cases. In the United States, the past 12 month and lifetime prevalence of MDD among adults were 10.4% and 20.6%, respectively. Only 12% of patients with MDD were hospitalized at some point in their life [36]. This means that 10–15% of true MDD cases are among MDD hospitalized group and 80–85% of them are among the non-exposed group. In other words, on average 8% of the patients who are hospitalized due to other reasons are affected by MDD. In Iran, data from the national mental health survey (IranMHS) showed that 12.7% of the adult population had suffered from MDD in the past twelve months [20], 41.2% of them had received any health services for their mental problem [37]. This means that if we use electronic health records, about 40% of adults with past twelve MDD will be found, and the rest 60% will retain in unexposed or healthy group. This non-differential misclassification of exposure will cause underestimation of the investigated relationship.
Second, all of the previous studies, except the UK biobank cohort study, were retrospective cohorts and have the typical limitations in measurement and adjustment of confounding factors. In our study, all of the confounding variables such as comorbidities, socioeconomic status, smoking, and obesity were carefully measured in cohort baseline assessments.
Third, in this study we have two annual follow-ups for all of the subjects with and without MDD. By far, this has been the most rigorous follow-up for assessing the study hypothesis.
One of the limitations of our study is its generalizability. The majority of our study population were working at hospitals and public health centers and almost half of them were healthcare professional. Therefore, the findings should be cautiously generalized to all of the employees or the general population. As we discussed earlier, this limitation might be the reason of no association found between MDD and COVID-19 hospitalization.
Conclusion. The findings of our study suggest that preexisting MDD is associated with higher susceptibility to COVID-19 infection but it has no significant relationship with hospitalization. Therefore, there is a need to identify and address modifiable risk factors and prevent delays in providing health care in this population. Future research should address whether COVID-19 vaccination has differential efficacy in those with MDD and whether infection with COVID-19 can change the natural course of MDD over time.