In the present study, using data from a naturalistic prospective pharmacotherapeutic study, which reflects real-world clinical practice, significant individual effects of baseline serum 5-HT level and age on 12-week remission were not found. Both a higher 5-HT level and age ≥ 60 years were associated with the highest 12-week remission rates with a significant multiplicative interaction effect and remained significant even after adjustment for relevant covariates and regardless of the initial antidepressant type.
As stated in the Introduction, the associations between peripheral (plasma, serum, and platelet) 5-HT levels and antidepressant treatment outcomes have been inconsistent in previous clinical studies. Reportedly, high baseline plasma 5-HT levels were associated with better 4- and 8-week SSRI responses in several studies15,16, indicating that treatment-related changes in plasma 5-HT levels, as well as baseline plasma 5-HT levels, are associated with a better SSRI response. However, in other studies, high baseline platelet 5-HT levels were associated with worse 4- and 6-week SSRI responses17,18. In another study, baseline serum or plasma 5-HT levels exhibited no association with the 4-week SSRI response19. The discrepancies in the results of previous studies do not appear to be due to differences in study design. The treatment response was assessed around 4 to 8 weeks in all studies. Furthermore, all studies included patients with major depressive disorders. Although the types of SSRIs prescribed were different, inconsistent results were also found in two studies with similar study designs in which the predictive effect of pre-treatment plasma 5-HT levels for the 4-week SSRI response was evaluated15,19.
In previous studies in which the predictive effect of baseline 5-HT levels for antidepressant treatment responses was investigated, the interaction between 5-HT levels and other variables was not considered. This may be the reason for the conflicting results regarding the association between baseline peripheral 5-HT levels and future antidepressant treatment responses. The present study results showing that a higher serum 5-HT level predicted a better treatment response only in patients ≥ 60 years of age may provide clues for addressing the current controversy. Synergistic effects between high 5-HT levels and age ≥ 60 years in our cohort are biologically plausible. Decreased central HTR binding activity, especially in HTR1A11,12,28 and HTR2A10,29, measured using positron emission tomography have been reported in patients with depressive disorders. Because HTR1A21–23 and HTR2A24–27 binding capacities are downregulated in older individuals, high 5-HT levels may be predictive of 12-week remission only in subjects ≥ 60 years of age due to decreased 5-HT signaling activity in the elderly.
Unlike in previous studies in which the association between 5-HT levels and antidepressant treatment outcomes in patients who only received SSRIs was analyzed15–19, we included patients who received various types of antidepressants. Notably, a multiplicative interaction effect of higher 5-HT level and age ≥ 60 years on 12-week remission was observed in patients who received initial SSRI treatment as well as in patients who received initial non-SSRI treatment. One potential mechanism underlying this observation is the change in tryptophan metabolism due to systemic inflammation, which shifts tryptophan metabolism from the 5-HT pathway to the kynurenine pathway by inducing the activity of indoleamine-2,3-dioxygenase30. Because increased inflammation is associated with a worse antidepressant response regardless of antidepressant type31, low 5-HT levels may be indicative of high inflammatory activity and therefore predict worse antidepressant responses regardless of antidepressant type. Another potential explanation for our result is decreased tryptophan availability. Tryptophan depletion reportedly decreases the mood in remitted major depressive disorder (MDD) patients regardless of the use of antidepressants32. Because a decreased mood can be a predictor of a worse antidepressant response33 and peripheral 5-HT levels are determined in part by tryptophan availability, we hypothesize that a lower 5-HT level may predict a worse antidepressant response by acting as a surrogate of low tryptophan availability regardless of antidepressant type.
However, several issues should be considered. First, because 5-HT cannot cross the blood-brain barrier, the peripheral and central 5-HT systems are functionally separated, and the correlation between peripheral and central 5-HT levels is controversial13,14,34. Further studies are needed to determine the detailed mechanisms by which peripheral 5-HT levels affect the antidepressant response in patients with depressive disorders. Second, we measured 5-HT levels in serum rather than in plasma or platelets. Peripheral 5-HT is mainly synthesized by enterochromaffin cells in the gut35. Once released from the gut, most 5-HT is taken up into platelets (> 95%), and the remaining free 5-HT level in the circulation is very low36–38. Therefore, plasma 5-HT levels reflect the bioactive free 5-HT, platelet 5-HT levels reflect the major 5-HT pool in the periphery, and serum 5-HT levels reflect both the bioactive 5-HT and 5-HT pools38,39. In the present study, we showed that the serum 5-HT level can be used as a biomarker of the antidepressant response. Based on these results, we hypothesize that both the bioactive 5-HT and 5-HT pools in the periphery may be predictive of antidepressant treatment outcomes in patients ≥ 60 years of age. To verify this hypothesis, the interaction effect of plasma and/or platelet 5-HT levels and age ≥ 60 years on 12-week remission should be investigated in patients with depressive disorders.
The strength of this study was the large sample size, and participants were evaluated using a structured research protocol and standardized scales. As stated above, this is the first study in which the interaction effect of peripheral 5-HT level and age on antidepressant treatment outcome was reported. Furthermore, because this was a naturalistic prospective study, which reflects actual clinical practice situations, results obtained in this study can serve as a basis for developing biomarkers of antidepressant treatment response in real-world clinical practice.
The present study had several limitations. First, the association between treatment-related changes in serum 5-HT levels and treatment outcomes could not be assessed because longitudinal data on serum 5-HT levels were lacking. Second, because the study was naturalistic in design, treatment was based on patient preferences under the guidance of a physician rather than determined based on a preset protocol; therefore, inter-physician variability might have affected the outcomes. However, because physicians guided treatment decisions without knowing the baseline 5-HT levels, inter-physician variability likely did not affect the outcomes. Third, because various antidepressants were initially used and different treatment strategies (S, A, C, and mixtures of these approaches) were implemented from 3 weeks after the start of antidepressant monotherapy, too many variables existed for evaluating the interaction effect of 5-HT level and age on the antidepressant response based on the type of antidepressant.
Despite the recognized antidepressant role of 5-HT/HTR signaling pathways in the CNS, the association between baseline peripheral 5-HT level and the antidepressant treatment response in clinical studies remains debatable. In the present study, the predictive value of the interaction between a higher 5-HT level and age ≥ 60 years for 12-week remission in patients with depressive disorders who received stepwise antidepressant therapy was investigated. Based on the results, we suggest that the combination of baseline serum 5-HT level and age can be a useful predictor of 12-week antidepressant treatment outcomes. Because this was a non-randomized trial, the interaction effect of serum 5-HT level and age as a predictor of treatment outcome in patients receiving antidepressants should be evaluated in future randomized trials.