Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Modifiable risk factors of major depressive disorder: A Mendelian randomization study

Abstract

Identifying modifiable risk factors early on is essential to prevent major depressive disorder (MDD). This study systematically investigated the causal relationship between 19 modifiable risk factors and MDD. Single-nucleotide polymorphisms (SNPs) associated with 19 potentially modifiable risk factors were screened via the genome-wide association study (GWAS) enrolling individuals of European descent. Summary-level data for MDD (59,851 cases and 113,154 controls) were extracted from the UK Biobank. The inverse-variance-weighted (IVW) method was utilized as the primary analysis. Sensitivity analyses were performed using the MR-Egger method, the Maximum likelihood method, the MR-pleiotropy residual sum outlier (MR-PRESSO) method, and MR-robust adjusted profile score (MR-RAPS) method. MR-Egger regression, heterogeneity tests, pleiotropy tests, and leave-one-out tests were also performed to analyze sensitivity. The MR Steiger test was used to verify the directionality of the exposure to the outcome. Genetically predicted smoking initiation increased the risk of MDD (P = 6.00E-09), while smoking status: never and past tobacco smoking decreased the risk of MDD (all P < 0.01). In addition, education level was inversely associated with MDD risk (all P < 0.01). Genetically instrumented sleeplessness/insomnia, daytime naps, and nap during the day were positively related to the risk of MDD (all P < 0.01). Personal feelings, including guilt, hurt, tension, and worry too long after an embarrassing experience, had a suggestive increased risk for MDD (all P < 0.000). The remaining five modifiable risk factors were all causally associated with the risk of MDD, including neuroticism, neuroticism scores, body mass index (BMI), average total household income before tax, and types of physical activity in the last 4 weeks (all P < 0.01). All 19 potentially modifiable risk factors were causally associated with the risk of MDD. The main hypothesis of this MR study was that identifying and intervening in these 19 potentially modifiable risk factors could be beneficial to the prevention and treatment of MDD and further reduce mortality and economic burden.

Citation: Zheng X, Tong L, Zhang C, Zhang C, Zhang C, Wan B (2023) Modifiable risk factors of major depressive disorder: A Mendelian randomization study. PLoS ONE 18(8): e0289419. https://doi.org/10.1371/journal.pone.0289419

Editor: Marianna Mazza, Universita Cattolica del Sacro Cuore Sede di Roma, ITALY

Received: February 26, 2023; Accepted: July 18, 2023; Published: August 3, 2023

Copyright: © 2023 Zheng et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: The RAW data is available in the Supplementary Files.

Funding: Hainan Province Clinical Medical Center (No. QWYH202175), the Scientific Research Project of Hainan Health Committee (No. 21A200115), the Research and Cultivation Fund of Hainan Medical University (No. HYPY2020015), the specific research fund of The Innovation Platform for Academicians of Hainan Province, the Natural Science Foundation of Hainan Province (No. 820RC771), and the Key R&D Projects of Hainan Province (No. ZDYF2022SHFZ074, No. ZDYF2022SHFZ280, No. ZDYF2017086, and No. ZDYF2019157). The funders had no role in study design, data collection, and analysis, the decision to publish, or the preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Introduction

Major depressive disorder (MDD) is a worldwide disease characterized by persistent low mood. MDD has an estimated prevalence of 4.4% worldwide, characterized by high mortality and disability rates [1–3]. In addition, MDD also imposes a substantial economic burden on the United States [4]. Compared with healthy people, patients with MDD have a higher risk for suicide and more serious medical problems, thus causing a huge social burden. MDD is influenced by multiple environmental and genetic factors, such as sexual, regional brain volumes and physical or emotional abuse during childhood, which are difficult to change [5]. Necessarily, further exploration of risk factors, which are easily controlled and prevented, for MDD can help design and implement effective prevention strategies or novel treatments.

Several personal and sociodemographic factors are related to MDD in traditional research, such as physical activity [6, 7], smoking [8, 9], education level [10], sleeping [11, 12], and BMI [13]. Cannabis has been, over the past two decades, one of the most widely used illegal psychoactive substances, particularly among adolescents and young adults. The association between cannabis use, mental illness, and suicidal behavior has been frequently reported, although this link’s exact nature is still poorly investigated. The paper’s results showed that Cannabis use was a relevant risk factor associated with suicidal attempts and behaviors in psychotic and non-psychotic samples [14]. In addition to depression, substance use disorders, particularly cannabis use, are recognized as significant risk factors for suicidal behavior in clinical and community populations [14]. Moreover, some epidemiological researchers suggest that personal feelings, including neuroticism items, may exacerbate MDD [15–22]. Besides, immune-inflammatory changes are important in the pathophysiology of major depression and suicidal behavior [23]. It is well known that traditional observation research is subjected to residual confounding factors and cannot establish a causal relationship. Therefore, elucidating the causal role of these controllable risk factors in MDD is very important to guide clinicians in preventing and treating MDD.

As a novel method, Mendelian randomization (MR), which treats heritable variation as a natural experiment, has been used to assess the potential causal association between multiple diseases [24]. Genetic variants are randomly allocated before birth, never change after birth, and unaffected by environmental factors; thus, consider making up for the lack of observation of study-residual confounding and reverse causation. Some MR studies have demonstrated the association between risk factors and MDD, but their exposure or outcome was single and not detailed enough [25–27]. Here, we demonstrate whether there is a causal relationship between 19 modifiable risk factors and MDD through the two-sample MR method using the latest and more extensive data from IEU Open GWAS database.

Materials and methods

Assumption of Mendelian randomization study

The MR research is based on the following principles:(i) Genetic instrument variables (GIVs) are significantly associated with exposures (19 potentially modifiable risk factors). The P value (P < 5 × 10−8) of the GIVs and the clumping method to exclude linkage disequilibrium (LD) were used to ensure the above assumption; (ii) GIVs are independent of any residual confounding factors; (iii) The affection of GIVs on the outcome (MDD) only through risk factors without horizontal pleiotropy [28, 29] (Fig 1). 19 controllable risk factors of our study were divided into five sections: smoking-related phenotypes, education-related phenotypes, sleeping-related phenotypes, personal feeling-related phenotypes, and other phenotypes.

thumbnail
Download:
Fig 1. Mendelian randomization (MR) model.

The MR research is based on the following principles:(i) Genetic instrument variables (GIVs) are significantly associated with exposures (19 potentially modifiable risk factors); (ii) GIVs are independent of any residual confounding factors; (iii) The affection of GIVs on the outcome only through risk factors without horizontal pleiotropy.

https://doi.org/10.1371/journal.pone.0289419.g001

Exposure and outcome data

The data supporting this study’s findings are openly available in IEU Open GWAS database at https://gwas.mrcieu.ac.uk/.

Based on wide phenotypic screening, modifiable and significant risk factors were selected for inclusion in the study. 19 potential modifiable risk factors used in our study were classified into five categories: (i) smoking-related phenotypes, including smoking initiation, past tobacco smoking, and smoking status: never; (ii) education-related phenotypes, including years of schooling, qualifications: college or university degree, qualifications: A levels/AS levels or equivalent and age completed full-time education; (iii) sleeping-related phenotypes, including sleeplessness/insomnia, daytime nap, and nap during day; (iv) personal feeling-related phenotypes, including feeling guilty, worry too long after an embarrassing experience, feeling hurt and feeling tense; (v) other phenotypes, including neuroticism, neuroticisms scores, body mass index (BMI), average total housed income before tax and types of physical activity in last 4 weeks: heavy DIY (e.g., weeding, lawn mowing, carpentry, digging) (S1 Table).

The outcome data was MDD (59,851 cases vs. 113,154 controls, ieu-1188). Participants in the 19 potentially modifiable risk factors were not screened for MDD.

Two-sample MR

SNP selection.

Based on the GWAS summary-level data of 19 potentially modifiable risk factors, SNPs with a lower P value (P < 5 × 10–8), longer physical distance (≥ 10,000 kb), and a lower likelihood of LD (r2 < 0.001) were maintained. Data of all GIVs used in this work are shown in S2 Table. The F statistic estimating the instrument strength of each SNP in a two-sample MR study was obtained from the published article [30, 31].

Primary analyses.

The multiplicative random effects inverse-variance-weighted (IVW) as the primary method was used to calculate the combined effects of all SNPs [32].

Sensitivity analyses.

We also performed the MR-Egger method [33], the maximum likelihood method [34], the MR-pleiotropy residual sum outlier method(MR-PRESSO) method [35], and the MR-robust adjusted profile score (MR-RAPS) [36] to evaluate the reliability and stability of the results. The IVW and MR-Egger methods were performed to estimate the heterogeneity, where the P value greater than 0.05 indicated no heterogeneity. The MR-PRESSO, MR-Egger, and IVW methods were employed to identify and remove outliers. The intercept of the MR-Egger model was used to test the pleiotropy, where a deviation from 0 denotes the presence of directional pleiotropy. What’s more, the leave-one-out method was implemented to assess the influence of a single SNP on the total effect of IVW. An available website MR method, referenced in the published article, was used to calculate the statistical power of MR analysis, where a power greater than 80% was considered an excellent value [37]. When the P value of the MR Steiget test was lower than 0.05, the causal direction of exposure-causing the outcome was statistically significant [38].

All the MR analyses were implemented via the TwoSampleMR [39] and MRPRESSO packages [35] in R software (V 4.2.2).

Results

Baseline information

The essential features of the 19 potentially modifiable risk factors are summarized in S1 Table. 9–262 SNPs were associated with the 19 modifiable risk factors (all P < 5 × 10−8). In addition, there was no weak instrument bias in the two-sample MR study (all F statistics > 10) (Table 1).

thumbnail
Download:
Table 1. Results of the MR analysis of the 19 modifiable risk factors for MDD.

MR, Mendelian randomization; SNPs, single nucleotide polymorphisms; OR, odds ratio; GWAS, genome-wide association studies; IVW, inverse-variance-weighted; MR-PRESSO, MR-pleiotropy residual sum outlier; MR-RAPS, MR-robust adjusted profile score.

https://doi.org/10.1371/journal.pone.0289419.t001

Otherwise, the statistical power for outcome MDD was approximately equal to 100% (Table 1). S2 Table summarizes the details of all SNPs (S2 Table).

No evidence of directional pleiotropy was found based on the intercept of MR-Egger. (Table 1). The results of the leave-one-out analysis showed that no single SNP influenced the overall estimates (S3 Table). The influence of the modifiable risk factors on MDD risk was the correct causal direction by the MR Steiger test (all P < 0.01) (Table 1).

Smoking-related phenotypes for the risk of MDD

As for the smoking-related phenotypes in our MR study, smoking initiation tends to increase the risk of genetically developing MDD. Under the IVW method, the odds ratio (OR) (95% CI) was 1.386 (1.225–1.568; P = 6.00E-09) (Table 1). Other MR methods, including Maximum likelihood, MR-RAPS, and MR-PRESSO, showed the same sensitivity (Table 1, S1A and S2A Figs).

In addition, smoking status: never and past tobacco smoking were all inversely and genetically related to MDD, where the OR was 0.342 (0.238–0.491; P = 6.61E-09) and 0.723 (0.627–0.834; P = 8.16E-06) via the IVW method (Table 1). More importantly, the Maximum likelihood, MR-RAPS, and MR-PRESSO methods significantly showed the same sensitivity, indicating the results were reliable and stable (Table 1, S1B, S1C, S2B and S2C Figs).

Education-related phenotypes for the risk of MDD

Regarding the education-related phenotypes, they all genetically and protectively decreased the risk of MDD. The OR was 0.693 (0.620–0.776; P = 1.61E-10), 0.542 (0.431–0.682; P = 1.66E-07), 0.414 (0.265–0.648; P = 1.15E-04) and 0.631 (0.468–0.850; P = 2.48E-03), respectively (Table 1). We found the same protective association via the Maximum likelihood, MR-PRESSO, and MR-RAPS methods, indicating the results were reliable and stable (Table 1, S1D–S1G, S2D–S2G Figs).

Sleeping-related phenotypes for the risk of MDD

Genetic predisposition to sleeping-related phenotypes was significantly related to MDD under the IVW method (Table 1). Those were supported by the Maximum Likelihood, MR-PRESSO, and MR-RAPS methods, although the MR-Egger method did not reach statistical significance (Table 1, S1H–S1J, S2H–S2J Figs).

Feeling-related phenotypes for the risk of MDD

To our primary MR method, the IVW method genetically showed that personal feeling-related phenotypes were all associated with the overall risk of advanced MDD (Table 1). Maximum Likelihood, MR-PREEO method, and MR-RAPS method for the association of feeling-related risk factors with outcome MDD were broadly similar to our primary MR method (Table 1, S1K–S1N, S2K–S2N Figs), indicating that our results were reliable and stable.

Other phenotypes for the risk of MDD

As for the remaining phenotypes, there were suggestive associations between three risk factors (neuroticism, neuroticism scores, and BMI) and MDD. The results of the IVW method are shown in Table 1. Other MR methods used in our study showed the same sensitivity (Table 1, S1O–S1Q, S2O–S2Q Figs).

Regarding average total household income before tax and types of physical activity in the last 4 weeks: heavy DIY (e.g., weeding, lawn mowing, carpentry, digging), we found an inverse association between the two risk factors and MDD under the IVW method (OR = 0.679, P = 1.71E-05; OR = 0.402, P = 1.07E-02). The sensitivity of Maximum likelihood, MR-PREEO, and MR-RAPS methods for the association of the two factors with MDD were broadly similar to the IVW method (Table 1, S1R, S1S, S2R and S2S Figs), indicating that our results were reliable.

Discussion

Genetically, the results suggested that smoking, low education level, poor sleep quality, and negative personal emotion were all positively and significantly related to the risk of MDD. Moreover, among the 19 personal and sociodemographic controllable risk factors, the remaining factors had significant casual relationships with MDD. The main hypothesis of this MR study was that identifying and intervening in these 19 potentially modifiable risk factors could be beneficial to the prevention and treatment of MDD and further reduce mortality and economic burden.MDD is a complex, pervasive and burdensome disorder with links to multiple factors [40], including smoking behaviour [41]. The results from some traditional researches have reported that smoking-related phenotypes are related to MDD [8, 9, 42], agreeing with the results of our two-sample MR study assessing the causal direction (P < 0.01). Conversely, a recent MR study genetically predicated smoking as a protective factor for MDD [43]. They just analyzed the pregnant women, while our outcome data consisted of men and women and was more representative. Some MR studies also explored the relationship between smoking and MDD, but their exposure or outcome was single and not detailed enough [25–27]. On the contrary, we studied multiple smoking phenotypes (smoking initiation, past tobacco smoking, and smoking status: never smoking) and MDD, one of the sub-categories of depression, and the data of our MR study is the latest release and more extensive than others. Genetically, our findings remind the public that reducing or not smoking may reduce the risk of MDD.

Epidemiological and observational investigation and review articles have repeatedly explored the relationship between sleep-related phenotypes and MDD [11, 12, 44–47]. Although some investigators accounted for confounding factors and repeated participants, residual confounding and reverse causality are still unavoidable in epidemiological and observational investigations [48]. Therefore, the MR analyses seem critical in addressing the above problems. Our MR study genetically predicted that sleeplessness/insomnia, nap during the day, and daytime nap aggravated MDD, which indicated that improved sleep quality was beneficial for preventing and treating MDD. Another two MR studies genetically reported that sleep-related phenotypes were not associated with MDD [49, 50]. It should be pointed out that the sample size of exposure and outcome of our MR study was more significant than the above MR study, so our results are more credible, making the public more aware of the importance of increasing sleep quality in reducing the risk of MDD.

Educational level is a crucial sociodemographic determinant of health and influences many diseases and disorders. In a Chinese study, Gan and colleagues discovered that education level was positively related to the risk of MDD, regardless of men or other races [51]. Their results did not support the straightforward claim made in publications from Europe and the US that higher levels of education were linked to reduced prevalence of MDD [52]. Reports from Finland showed no discernible differences across various educational groupings [53]. Another study linked illiteracy among older adults with higher incidence and more severe depression [54]. The above evidence for conflicting results of the effect of education on MDD came from observational studies and tended to suffer from residual confounding. We explored the genetic association between different education-related phenotypes and the risk of MDD. The results suggested that increasing personal and educational levels may decrease the risk of MDD, making the public aware of the importance of raising the education level and living to learn.

Guilty is one of the human emotions and is fascinating and complex. Feeling guilty has negative and positive health-related outcomes and incredibly contradictory empirical results in studies of depression [22, 55, 56]. Some research posited guilty as a positive, healthy construct [57, 58]. Conversely, some researchers reported that feeling guilty was significantly associated with major depression [22, 59–64]. However, the casual direction was not revealed and suffered from residual confounding. We were the first to evaluate the relationship between guilty and MDD using the mendelian randomization method. We genetically suggested that by using several MR methods, feeling guilty was related to a higher risk for MDD. The trauma experienced by people with severe mental diseases extends beyond physical and sexual abuse. The results in a group of female patients having urodynamics revealed that depression was linked to more severe shame. [21]. One interpretive phenomenological research concluded that MDD patients had increased relative frequencies of embarrassment. [65].

On the contrary, Louis et al. reported that depressive symptoms might amplify embarrassment in essential tremors [66]. Our study showed that worrying too long after an embarrassing experience was related to an increased risk for MDD. Fewer studies explored whether feeling tense and hurt concerning MDD, and the results were equivocal and influenced by other factors [15–20]. Our work showed that genetically predicated feeling hurt and tense was associated with an increased risk for MDD. All the above findings about personal emotions remind the public to pay attention to controlling their emotions in daily life.

A precursor to mental illness, neuroticism is a propensity for unpleasant feelings such as melancholy, worry, and emotional instability. Many studies showed that high neuroticism was a strong vulnerability marker for developing MDD and was common among those with MDD [67–70]. In an MR study, they found neuroticism and its items were genetically associated with the risk for MDD, but they did not show the statistical power and the causal direction. As for our MR study, we found neuroticism, its items (feeling guilty, worrying too long after an embarrassing experience, feeling hurt), and neuroticism scores were connected to a higher risk of MDD. Additionally, the MR Steiger test confirmed the casual assumption that neuroticism or its items correlate with MDD, and the outcome supported the idea. (P < 0.01)). The power of our investigation was around 1. Moreover, the exposure data were from different sources, illustrating the association between Neuroticism and MDD, and the result was more convincing. The significant conclusion of our results is that lowering neuroticism would be advantageous in reducing the risk of MDD.

The growing body of research showing a link between physical activity and a lower risk of depression offers a potentially versatile option for prevention.[6, 7, 71]. Uncertainty exists over the association’s direction and causation, though. According to an MR investigation, Accelerometer-based exercise appeared to protect against MDD. Self-reported activity and MDD did not significantly relate, indicating that different physical activities may impact MDD [72]. Another MR study genetically suggested that physical activity is prospectively associated with depression [73]. Regarding our MR study, different physical activity from the above exposure was genetically associated with the reduced risk for MDD, which further demonstrated that physical activity is essential in lowering MDD and reminded people to increase DIY activity to prevent MDD. Regarding BMI, our MR findings revealed an inverse relationship between BMI and MDD, indicating that the general public should maintain a healthy BMI to lower their chance of developing MDD.

The primary significance of our research is that we first conclude 19 modifiable risk factors and divide them into five categories, which are genetically associated with the risk for MDD, with stringent quality control conditions and analysis methodologies and produce dependable and steady research results. Unlike other MR studies, the risk factors of our MR study are more detailed and multifaceted, making the results more reliable. In addition, the sample size of our research is larger and latest than other studies. The results benefit the prevention and treatment of MDD by paying attention to the 19 risk factors.

We must admit that our study has some flaws. The results of our study may not apply to people of different origins because we only included participants of European ancestry. Thus, it is important to proceed with caution when interpreting the implications of our findings. Second, it would be intriguing to determine SNPs linked to the severity of MDD and analyses the relationship between modifiable risk variables and MDD severity. Since there are no publications on the associations between certain SNPs and the severity of MDD, we cannot do this association analysis. Thirdly, there is variability in the instrumental variables, possibly due to real causality rather than a failure to uphold the instrument variable premise. As for heterogeneity, we used the IVW as the primary analysis method. Fourth, other factors, such as other diseases, may confound our MR results, but avoiding them isn’t easy.

Conclusions

There is a causal relationship between 19 modifiable risk factors and MDD through the two-sample MR method using the latest and more extensive data from IEU Open GWAS database, and identifying and intervening in these 19 potentially modifiable risk factors could be beneficial to the prevention and treatment of MDD and further reduce mortality and economic burden.

Supporting information

S1 Fig. Density plots of the MR study.

Density plot of the MR results of 19 modifiable risk factors for MDD. Represent the results of heterogeneity analysis from 19 modifiable risk factors. MR, Mendelian randomization; MDD, major depressive disorder; SNP, single nucleotide polymorphisms; IVW, inverse-variance-weighted; MR-PRESSO, MR-pleiotropy residual sum outlier; MR-RAPS, MR-robust adjusted profile score.

https://doi.org/10.1371/journal.pone.0289419.s001

(TIF)

S2 Fig. Scatter plots of the MR study.

On a log-odds scale, a scatter plot shows how each risk factor-related SNP affects MDD. MR, Mendelian randomization; MDD, major depressive disorder; SNP, single nucleotide polymorphisms; IVW, inverse-variance-weighted; MR-PRESSO, MR-pleiotropy residual sum outlier; MR-RAPS, MR-robust adjusted profile score.

https://doi.org/10.1371/journal.pone.0289419.s002

(TIF)

S1 File. Raw data.

https://doi.org/10.1371/journal.pone.0289419.s003

(XLSX)

S1 Table. Data characteristics from the GWAS summary.

SNP = single nucleotide polymorphism; GWAS, genome-wide association studies.

https://doi.org/10.1371/journal.pone.0289419.s004

(DOCX)

S2 Table. SNPs strongly associated with 19 risk factors and their F statistic.

SNPs, single nucleotide polymorphisms; EAF, Effect allele frequency.

https://doi.org/10.1371/journal.pone.0289419.s005

(DOCX)

S3 Table. Leave-one-out results of the MR study.

SNPs, single nucleotide polymorphisms.

https://doi.org/10.1371/journal.pone.0289419.s006

(DOCX)

References

  1. 1. Liu Q., et al., Changes in the global burden of depression from 1990 to 2017: Findings from the Global Burden of Disease study. J Psychiatr Res, 2020. 126: p. 134–140. pmid:31439359
  2. 2. Friedrich M.J., Depression Is the Leading Cause of Disability Around the World. JAMA, 2017. 317(15): p. 1517. pmid:28418490
  3. 3. Disease G.B.D., Injury I., and Prevalence C., Global, regional, and national incidence, prevalence, and years lived with disability for 354 diseases and injuries for 195 countries and territories, 1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. Lancet, 2018. 392(10159): p. 1789–1858. pmid:30496104
  4. 4. Greenberg P.E., et al., The Economic Burden of Adults with Major Depressive Disorder in the United States (2010 and 2018). Pharmacoeconomics, 2021. 39(6): p. 653–665. pmid:33950419
  5. 5. Otte C., et al., Major depressive disorder. Nat Rev Dis Primers, 2016. 2: p. 16065. pmid:27629598
  6. 6. Philippot A., et al., Impact of physical exercise on depression and anxiety in adolescent inpatients: A randomized controlled trial. J Affect Disord, 2022. 301: p. 145–153. pmid:35007642
  7. 7. Pearce M., et al., Association Between Physical Activity and Risk of Depression: A Systematic Review and Meta-analysis. JAMA Psychiatry, 2022. 79(6): p. 550–559. pmid:35416941
  8. 8. Vigo D., Thornicroft G., and Atun R., Estimating the true global burden of mental illness. Lancet Psychiatry, 2016. 3(2): p. 171–8. pmid:26851330
  9. 9. Talati A., Keyes K.M., and Hasin D.S., Changing relationships between smoking and psychiatric disorders across twentieth century birth cohorts: clinical and research implications. Mol Psychiatry, 2016. 21(4): p. 464–71. pmid:26809837
  10. 10. Peyrot W.J., et al., The association between lower educational attainment and depression owing to shared genetic effects? Results in ~25,000 subjects. Mol Psychiatry, 2015. 20(6): p. 735–43. pmid:25917368
  11. 11. Roberts R.E. and Duong H.T., The prospective association between sleep deprivation and depression among adolescents. Sleep, 2014. 37(2): p. 239–44. pmid:24497652
  12. 12. Ohayon M.M., Caulet M., and Lemoine P., Comorbidity of mental and insomnia disorders in the general population. Compr Psychiatry, 1998. 39(4): p. 185–97. pmid:9675502
  13. 13. Dreimuller N., et al., Body mass index (BMI) in major depressive disorder and its effects on depressive symptomatology and antidepressant response. J Affect Disord, 2019. 256: p. 524–531. pmid:31280076
  14. 14. Serafini G., et al., Can cannabis increase the suicide risk in psychosis? A critical review. Curr Pharm Des, 2012. 18(32): p. 5165–87. pmid:22716157
  15. 15. Chu H.C. and Chow P.C., [Providing grief counseling to a major depressive elderly widower: a nurse’s experience]. Hu Li Za Zhi, 2008. 55(5): p. 90–6. pmid:18836981
  16. 16. Rief W. and Auer C., Is somatization a habituation disorder? Physiological reactivity in somatization syndrome. Psychiatry Res, 2001. 101(1): p. 63–74. pmid:11223121
  17. 17. Caldwell W.G. and Nowers W., Prepartum depression; a study of seventy unwed mothers treated with d-amphetamine sulfate and amobarbital. Calif Med, 1958. 88(5): p. 380–2. pmid:13523417
  18. 18. Persad S.M. and Polivy J., Differences between depressed and nondepressed individuals in the recognition of and response to facial emotional cues. J Abnorm Psychol, 1993. 102(3): p. 358–68. pmid:8408947
  19. 19. Piccirilli A.M. and Pos A.E., Does emotional processing predict 18-month post-therapy outcomes in the experiential treatment of major depression? Psychother Res, 2022: p. 1–13. pmid:35594384
  20. 20. Mucci N., et al., Predictors of trauma in bank employee robbery victims. Neuropsychiatr Dis Treat, 2015. 11: p. 2605–12. pmid:26504393
  21. 21. Yeung J.Y., Eschenbacher M.A., and Pauls R.N., Pain and embarrassment associated with urodynamic testing in women. Int Urogynecol J, 2014. 25(5): p. 645–50. pmid:24280994
  22. 22. Luck T. and Luck-Sikorski C., Feelings of guilt in the general adult population: prevalence, intensity and association with depression. Psychol Health Med, 2021. 26(9): p. 1143–1153. pmid:33295786
  23. 23. Serafini G., et al., A Specific Inflammatory Profile Underlying Suicide Risk? Systematic Review of the Main Literature Findings. Int J Environ Res Public Health, 2020. 17(7). pmid:32244611
  24. 24. Byrne E.M., Yang J., and Wray N.R., Inference in Psychiatry via 2-Sample Mendelian Randomization-From Association to Causal Pathway? JAMA Psychiatry, 2017. 74(12): p. 1191–1192. pmid:29094155
  25. 25. Yuan S., Yao H., and Larsson S.C., Associations of cigarette smoking with psychiatric disorders: evidence from a two-sample Mendelian randomization study. Sci Rep, 2020. 10(1): p. 13807. pmid:32796876
  26. 26. Wootton R.E., et al., Evidence for causal effects of lifetime smoking on risk for depression and schizophrenia: a Mendelian randomisation study. Psychol Med, 2020. 50(14): p. 2435–2443. pmid:31689377
  27. 27. Cai J., et al., Socioeconomic status, individual behaviors and risk for mental disorders: A Mendelian randomization study. Eur Psychiatry, 2022. 65(1): p. e28. pmid:35431011
  28. 28. Davey Smith G. and Hemani G., Mendelian randomization: genetic anchors for causal inference in epidemiological studies. Hum Mol Genet, 2014. 23(R1): p. R89–98. pmid:25064373
  29. 29. Smith G.D. and Ebrahim S., ’Mendelian randomization’: can genetic epidemiology contribute to understanding environmental determinants of disease? Int J Epidemiol, 2003. 32(1): p. 1–22. pmid:12689998
  30. 30. Wu F., et al., Mendelian randomization study of inflammatory bowel disease and bone mineral density. BMC Med, 2020. 18(1): p. 312. pmid:33167994
  31. 31. Pierce B.L., Ahsan H., and Vanderweele T.J., Power and instrument strength requirements for Mendelian randomization studies using multiple genetic variants. Int J Epidemiol, 2011. 40(3): p. 740–52. pmid:20813862
  32. 32. Burgess S., Butterworth A., and Thompson S.G., Mendelian randomization analysis with multiple genetic variants using summarized data. Genet Epidemiol, 2013. 37(7): p. 658–65. pmid:24114802
  33. 33. Bowden J., Davey Smith G., and Burgess S., Mendelian randomization with invalid instruments: effect estimation and bias detection through Egger regression. Int J Epidemiol, 2015. 44(2): p. 512–25. pmid:26050253
  34. 34. Xue H., Shen X., and Pan W., Constrained maximum likelihood-based Mendelian randomization robust to both correlated and uncorrelated pleiotropic effects. Am J Hum Genet, 2021. 108(7): p. 1251–1269. pmid:34214446
  35. 35. Verbanck M., et al., Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nat Genet, 2018. 50(5): p. 693–698. pmid:29686387
  36. 36. Qingyuan Z., et al., Statistical inference in two-sample summary-data Mendelian randomization using robust adjusted profile score. The Annals of Statistics, 2020. 48(3): p. 1742–1769.
  37. 37. Brion M.J., Shakhbazov K., and Visscher P.M., Calculating statistical power in Mendelian randomization studies. Int J Epidemiol, 2013. 42(5): p. 1497–501. pmid:24159078
  38. 38. Hemani G., Tilling K., and Davey Smith G., Orienting the causal relationship between imprecisely measured traits using GWAS summary data. PLoS Genet, 2017. 13(11): p. e1007081. pmid:29149188
  39. 39. Hemani G., et al., The MR-Base platform supports systematic causal inference across the human phenome. Elife, 2018. 7. pmid:29846171
  40. 40. Kendall K.M., et al., The genetic basis of major depression. Psychol Med, 2021. 51(13): p. 2217–2230. pmid:33682643
  41. 41. Fluharty M., et al., The Association of Cigarette Smoking With Depression and Anxiety: A Systematic Review. Nicotine Tob Res, 2017. 19(1): p. 3–13. pmid:27199385
  42. 42. An R. and Xiang X., Smoking, heavy drinking, and depression among U.S. middle-aged and older adults. Prev Med, 2015. 81: p. 295–302. pmid:26436684
  43. 43. Lewis S.J., et al., Smoking is associated with, but does not cause, depressed mood in pregnancy—a mendelian randomization study. PLoS One, 2011. 6(7): p. e21689. pmid:21818261
  44. 44. Riemann D., et al., Sleep, insomnia, and depression. Neuropsychopharmacology, 2020. 45(1): p. 74–89. pmid:31071719
  45. 45. Plante D.T., The Evolving Nexus of Sleep and Depression. Am J Psychiatry, 2021. 178(10): p. 896–902. pmid:34592843
  46. 46. Okun M.L., et al., Poor sleep quality increases symptoms of depression and anxiety in postpartum women. J Behav Med, 2018. 41(5): p. 703–710. pmid:30030650
  47. 47. Nematolahi P., et al., Effects of Rosmarinus officinalis L. on memory performance, anxiety, depression, and sleep quality in university students: A randomized clinical trial. Complement Ther Clin Pract, 2018. 30: p. 24–28. pmid:29389474
  48. 48. Daghlas I., et al., Genetically Proxied Diurnal Preference, Sleep Timing, and Risk of Major Depressive Disorder. JAMA Psychiatry, 2021. 78(8): p. 903–910. pmid:34037671
  49. 49. Huang J., et al., Sleep, major depressive disorder, and Alzheimer disease: A Mendelian randomization study. Neurology, 2020. 95(14): p. e1963–e1970. pmid:32817390
  50. 50. Gao X., et al., The bidirectional causal relationships of insomnia with five major psychiatric disorders: A Mendelian randomization study. Eur Psychiatry, 2019. 60: p. 79–85. pmid:31234011
  51. 51. Gan Z., et al., The impact of educational status on the clinical features of major depressive disorder among Chinese women. J Affect Disord, 2012. 136(3): p. 988–92. pmid:21824664
  52. 52. Alonso J., et al., 12-Month comorbidity patterns and associated factors in Europe: results from the European Study of the Epidemiology of Mental Disorders (ESEMeD) project. Acta Psychiatr Scand Suppl, 2004(420): p. 28–37. pmid:15128385
  53. 53. Lindeman S., et al., The 12-month prevalence and risk factors for major depressive episode in Finland: representative sample of 5993 adults. Acta Psychiatr Scand, 2000. 102(3): p. 178–84. pmid:11008852
  54. 54. Kim B.S., et al., Impact of illiteracy on depression symptomatology in community-dwelling older adults. Int Psychogeriatr, 2014. 26(10): p. 1669–78. pmid:24945628
  55. 55. Szentagotai-Tatar A. and Miu A.C., Correction: Individual Differences in Emotion Regulation, Childhood Trauma and Proneness to Shame and Guilt in Adolescence. PLoS One, 2017. 12(1): p. e0171151. pmid:28122023
  56. 56. Muris P., et al., Lack of guilt, guilt, and shame: a multi-informant study on the relations between self-conscious emotions and psychopathology in clinically referred children and adolescents. Eur Child Adolesc Psychiatry, 2016. 25(4): p. 383–96. pmid:26220659
  57. 57. Tangney J.P., Wagner P., and Gramzow R., Proneness to shame, proneness to guilt, and psychopathology. J Abnorm Psychol, 1992. 101(3): p. 469–78. pmid:1500604
  58. 58. Tangney J.P., Moral affect: the good, the bad, and the ugly. J Pers Soc Psychol, 1991. 61(4): p. 598–607. pmid:1960652
  59. 59. Wagner B., Hofmann L., and Grafiadeli R., The relationship between guilt, depression, prolonged grief, and posttraumatic stress symptoms after suicide bereavement. J Clin Psychol, 2021. 77(11): p. 2545–2558. pmid:34081784
  60. 60. Tangney J.P., Stuewig J., and Mashek D.J., Moral emotions and moral behavior. Annu Rev Psychol, 2007. 58: p. 345–72. pmid:16953797
  61. 61. Rubin D.A., et al., Predictors, prevention, and long-term prognosis of atrial fibrillation after coronary artery bypass graft operations. J Thorac Cardiovasc Surg, 1987. 94(3): p. 331–5. pmid:3306163
  62. 62. Misiolek-Marin A., et al., Influence of Burnout and Feelings of Guilt on Depression and Health in Anesthesiologists. Int J Environ Res Public Health, 2020. 17(24). pmid:33322399
  63. 63. LeBlanc N.J., et al., Shame, guilt, and pride after loss: Exploring the relationship between moral emotions and psychopathology in bereaved adults. J Affect Disord, 2020. 263: p. 405–412. pmid:31969271
  64. 64. Kim S., Thibodeau R., and Jorgensen R.S., Shame, guilt, and depressive symptoms: a meta-analytic review. Psychol Bull, 2011. 137(1): p. 68–96. pmid:21219057
  65. 65. Colic J., et al., Depersonalization and derealization in embarrassing social interactions: an experience sampling study in social phobia, major depression and controls. J Anxiety Disord, 2020. 70: p. 102189. pmid:32070861
  66. 66. Louis E.D., Cosentino S., and Huey E.D., Depressive symptoms can amplify embarrassment in essential tremor. J Clin Mov Disord, 2016. 3: p. 11. pmid:27429787
  67. 67. Nagel M., et al., Meta-analysis of genome-wide association studies for neuroticism in 449,484 individuals identifies novel genetic loci and pathways. Nat Genet, 2018. 50(7): p. 920–927. pmid:29942085
  68. 68. Klein D.N., Kotov R., and Bufferd S.J., Personality and depression: explanatory models and review of the evidence. Annu Rev Clin Psychol, 2011. 7: p. 269–95. pmid:21166535
  69. 69. Kendler K.S., et al., Personality and major depression: a Swedish longitudinal, population-based twin study. Arch Gen Psychiatry, 2006. 63(10): p. 1113–20. pmid:17015813
  70. 70. Bagby R.M., et al., Personality and depression. Can J Psychiatry, 2008. 53(1): p. 14–25. pmid:18286868
  71. 71. Liang Z.D., et al., Association between sedentary behavior, physical activity, and cardiovascular disease-related outcomes in adults-A meta-analysis and systematic review. Front Public Health, 2022. 10: p. 1018460. pmid:36339165
  72. 72. Choi K.W., et al., Assessment of Bidirectional Relationships Between Physical Activity and Depression Among Adults: A 2-Sample Mendelian Randomization Study. JAMA Psychiatry, 2019. 76(4): p. 399–408. pmid:30673066
  73. 73. Choi K.W., et al., An Exposure-Wide and Mendelian Randomization Approach to Identifying Modifiable Factors for the Prevention of Depression. Am J Psychiatry, 2020. 177(10): p. 944–954. pmid:32791893