As one of the two subcomponents of BMI, FMI showed a distinct relationship with a higher frequency of participants scoring in the depressive range on the BDI. When analyzing the subtypes of depressive symptoms separately no such relationship was found in the melancholic group. In accordance with our hypothesis, we found that the group with non-melancholic depressive symptoms showed a more pronounced relationship with body composition than those with melancholic depressive symptoms. More specifically, non-melancholic depressive symptoms were associated with higher FMI. This may suggest that the combined effect showing a relationship between high FMI and depressive symptoms was explained by the findings in the non-melancholic group. This is an important finding, as prior research has established the relationship between increased fat mass and depression. [12] Our findings suggest that this may be limited to non-melancholic depressive symptoms rather than overall depressive symptoms or melancholic depressive symptoms.
Furthermore, our findings suggest that body fat mass and lean mass have opposite relationships with the prevalence of depressive symptoms. The presence of a high FMI increased the likelihood of depressive symptoms, whereas the presence of a high LMI decreased the likelihood of depressive symptoms. Since neither of these variables can exist independently of each other it is important to remember that their opposite effects will moderate each other. This further emphasizes the importance of considering these factors both individually and simultaneously.
A variety of possible explanations for the body composition – depression relationship exists. Research suggests that there may be some genetic overlap between obesity and depressive symptoms, that could be driving both states. [18] Further it has been suggested that the psychological aspects of beauty standards may play a partial role in mediating the relationship between overweight and depressive symptomatology. [1] However, a study with a 12-month follow-up found that increases in fat mass or BMI were not related to any increase in depression. [42]
The most common of the explanations for the body composition – depression relationship is that of an underlying inflammatory effect. Various inflammatory markers, including CRP, have been used as measurements of inflammatory responses within the body [1], and hsCRP was used in the current study as well. The current study found that high FMI was associated with a higher hsCRP concentration. Considering the association found between FMI and non-melancholic symptoms, it is not surprising that we found higher hsCRP concentration in the non-melancholic group than in the non-depressed group. This is in line with prior research showing that non-melancholic depressive symptoms correlate with CRP levels [11], and that those having both obesity and metabolic syndrome have the highest levels. [2]
The pathologies of both obesity and depression have inflammatory components [1,14], with obesity exhibiting chronic low-grade inflammation. [1] Visceral adipose tissue has a particularly high production of pro-inflammatory cytokines that play a role in both obesity and depression. [13] This link between the two pathologies via immunological and inflammatory pathways has been suggested to be bidirectional and self-perpetuating, leading to a vicious cycle of the body composition – depression relationship being exaggerated over time. [1] Research has also been able to show that brain-derived neurotrophic factor, which is associated with obesity in humans, can be downregulated as a result of inflammation driven emotional changes in animals. [1] This provides a possible explanation of how the inflammatory pathway may work, but overall little is still known about the details of the inflammatory connection.
We did not detect any difference in hsCRP concentrations between the melancholic and non-melancholic groups. This may simply be due to the low power of the melancholic group. Melancholic depression has been shown to have lower inflammatory markers than non-melancholic depression [11], and may hence not be part of this obesity – depression relationship. The current thinking is that melancholic depression is not associated with inflammation, but rather with a change in the HPA-axis regulation. [11] If the thought that inflammation is what causes the changes in body composition is true, then this would explain why there is no association between melancholic depression and higher FMI. To accurately assess changes in the HPA-axis, multiple parameters need to be tested [2], making any such analysis extremely difficult. It has however been reported that cortisol levels are positively associated with melancholic depression. [11] Further, research involving Cushing’s syndrome has been able to show a causal relationship between cortisol and depression. [1]
It is of importance to note that some participants in the present study had comorbid diseases that could possibly have influenced our findings. Of note are especially cardiovascular disease, which was more prevalent in the high FMI group. Our findings indicate that group B (high fat mass and high lean mass) had the highest prevalence of cardiovascular disease, which is in agreement with previous research indicating that the combination of high FMI and high LMI is in fact predictive of development of diabetes. [43] This same body composition profile was also associated with the highest cardiometabolic risk. [43]
Both lower LMI and lower FMI show a relationship with lower fasting plasma glucose levels. However, glucose metabolism seems to be more related to FMI as indicated by higher glucose concentrations at 30 and 120 minutes after an OGTT. It has previously been shown that non-melancholic depression is more closely related to higher fasting glucose concentrations than melancholic depression is. [11,23] Here, however, we found no differences between the depressive subtypes for fasting glucose, but at 2 h after the OGTT the non-melancholic group showed higher glucose concentrations than the non-depressed group. Plasma glucose 2h after an OGTT has been shown to be a better predictor of mortality than fasting glucose. [44] Furthermore, impaired glucose metabolism is known to be a cardiovascular risk factor. [44,45] Since all diabetics were excluded from our study sample the glucose concentrations are within the normal range even though there is a difference between the groups. The values themselves do not infer a greater risk of cardiovascular disease than the general population, but the difference between the group can serve as an indication of some underlying difference in glucose regulation in the non-melancholic group. This is in line with the findings that non-melancholic depression is associated with body composition, as glucose metabolism is often altered in obesity, and both are associated with depression. [1]
In accordance with prior research [33] we showed that blood pressure has a main effect, with both higher LMI and FMI being related to higher blood pressure. This is true for both systolic and diastolic pressures. For the depressive subtypes we were able to show that the melancholic group has both lower systolic and diastolic blood pressure than either the non-depressed or the non-melancholic group. Higher blood pressure has been associated with both obesity and elevated glucose concentrations in depressed individuals. [10] Melancholic depressed individuals have previously been shown to have lower systolic blood pressure than non-depressed individuals. [11] This is all in line with our findings of the non-melancholic group having a stronger relationship with dysfunctional glucose metabolism and body composition.
Previous studies have reported sex-differences in the body composition – depression relationship [14], with women displaying a stronger relationship than men. [1,16,17] However, it has been suggested that the difference may be due to the sex-specific body compositions. [13] Ideally all factors could have been analyzed separately for both sexes, which would have allowed us to evaluate for differences between men and women. While we were not able to stratify the analyses according to gender due to limited power in the analyses, we standardized the body compositions factors by sex around their sex-specific means in order to effectively eliminate the sex-specific differences in body composition due to the differences in fat and lean mass distributions in men and women.
The current study has both strengths and limitations. Among the strengths are the extensively phenotyped participants, overall large sample size, and random selection of participants from the pool. The cross-sectional nature of the study is a limitation because it allows for no inferences regarding directionality or time-dependent causality. Other limitations are the presence of comorbidities among participants, and self-reported depressive symptoms rather than clinically diagnosed depressive disorders. The limited subgroup size prevented us from analyzing men and women separately, which would have been of benefit in gaining the most information possible. Ideally, we would also have had cortisol measurements for all the participants, but that information was not available. The age range could be viewed as a limitation in how these findings can be generalized to the general population, or as a strength since this minimizes the effect of aging on the results.
It would be of interest for further studies to build on these findings by introducing time as a factor. Some previous research has suggested that the body composition – depression relationship could be reinforced over time. [14] Depression has also been shown to have a relationship with long term body composition in some adolescents. [15] Another study showed that during a 12-month follow up increased BMI, visceral adiposity, or body fat did not correlate with increased depression. [42] Focusing on these same factors as the current study but in relation to lifetime body composition could be interesting. As we know birthweight and changes in weight over a lifetime can affect comorbidities differently. [26,27,46] It may be of value to know if rather than just body composition at a certain time, changes in body composition over the life course are associated with prevalence of one or the other type of depressive symptoms.
Depression is a treatable disease for which there are many effective treatment options. [47] One challenge is the heterogenous nature of depression, and the fact that currently the subtypes are only distinguished based on self-reported criteria rather than established biomarkers. [47] While depression has been increasing in prevalence, up to half of depressed individuals may be inadequately treated [47], which tells us about the need for a better understanding of this disease. Underlying pathophysiology of the depressive subtypes may be a factor in how depression could be treated more efficiently.
The novelty of the current study is that it provides more specific information differentiating between non-melancholic depressive symptoms and melancholic depressive symptoms and their relationships with FMI and LMI than previously available publications.