Findings from this longitudinal study are consistent with the results of prior cross-sectional studies suggesting that depression is cross-sectionally associated with higher hs-CRP levels in a population without evidence of a chronic life-threatening illness (e.g., cancer, or renal or heart failure). Among our sample of adults not reporting clinically significant depressive symptomatology, depression score appears to be independently and positively correlated to hs-CRP. After controlling for potentially confounding factors, analysis by gender revealed that this association persisted only among women. The disappearance of this relationship among men after controlling for potential confounders, suggests that factors such as BMI, age, and resting heart rate (a marker of physical fitness) may confound this association among men. These factors also appear to mediate the relationship among women, but to a lesser extent.
Previous studies suggest that some types of depression are associated with weight gain which in turn may exacerbate inflammation by inducing leptin expression and increasing synthesis of inflammatory cytokines by adipose tissue [25
]. Thus, the relationship between depression and hs-CRP may, at least partially, be mediated by weight gain. While our participants were, on average, not in the clinically depressed range, our findings suggest that the range of depressive symptomatology may be important, with low levels of depressive symptomatology having significance in these associations. The proposition that elevated depressive symptomatology leads to weight gain, and increase in body weight leads in turn to an increase in hs-CRP, would be a plausible explanation for our observation of the attenuation of the relationship between depression and inflammation among women, and the disappearance of the effect among men, after controlling for BMI.
The “psychoneuroimmune link” theory as an explanation of the relationship between the range of depressive symptomatology and hs-CRP can not be confirmed from the observations in this study; however, our findings support an independent association between depression scores and hs-CRP observed in women. On the other hand, among women, the observation of an independent and inverse relationship between depression scores and hs-CRP, over time, suggests that the relationship is complex and might be mediated mostly by behavior, since depression is related to changes in appetite and body weight, in either direction.
Our results are in contrast with reports from the NHANES III survey [10
], which found that the relationship between hs-CRP and depression is stronger among men than in women. However, those reports used major depression diagnosis in contrast to depressive symptom scores as used in this analysis. The association between hs-CRP and depression was not observed in a recent study from adolescents aged 13–16 years old [26
]. Differences in population characteristics and in the confounders controlled, also could account for some of the inconsistent results.
There are several strengths to our investigation. First, we collected detailed longitudinal data on a quarterly basis over the course of one year in various domains, including demographic, psychosocial, dietary, and physical activity. Second, our study controlled for potential confounding factors for the depression and hs-CRP relationship, which previous studies had not included. Third, the study population is relatively healthy; subjects with diabetes and other chronic diseases were excluded as these conditions have been associated with both depressive symptoms and hs-CRP levels and may thus confound the BDI and hs-CRP relationship. Fourth, our data came from a study of seasonal variation in blood lipids, therefore patients planning to use or are using lipid medications and hormone therapy were excluded. Consequently, any effect of using statins and female hormones on hs-CRP concentrations was eliminated. Finally, use of a continuous measure of depressive symptomatology (BDI scores) allowed analyses across the continuum of depressive symptom severity.
Our study also has potential limitations. The study sample had a reduced range of depression scores, which limits our capacity to draw conclusions for individuals in the clinical-depression range. Although, the finding of a positive association between depressive symptom score and hs-CRP among adults who were on average not clinically depressed is also of interest, because it suggests that the entire range of depressive-symptomatology may be associated with health risks. On the other hand, participants in this study were predominantly white, well-educated and employed therefore, caution should be taken when generalizing to populations that have different demographic characteristics or depression scores differing from the range found in this study. Finally, as with many studies that require a strong commitment on the part of the participants, selection bias is always possible.
In conclusion, multivariable longitudinal analyses suggest that an independent association between depression scores and hs-CRP is present among women, but not among men. It appears that obesity partially mediates this relationship. Other factors that were significantly and positively related to hs-CRP included body mass index, age, resting-heart rate, and concurrent minor infection/inflammation process. Further studies are required to elucidate the biological mechanisms of these associations and their implications.