Assessing the Assessment of Brain Serotonin Turnover

I would like to comment on several critical aspects of the January 2008 article of Barton and colleagues¹ entitled “Elevated Brain Serotonin Turnover in Patients With Depression: Effect of Genotype and Therapy.”

The authors indicate that measured internal jugular venoarterial (VA) differences in plasma levels of 5-hydroxyindoleacetic acid (5-HIAA, the major serotonin metabolite) are proportionally related to brain serotonin turnover. It is difficult to reconcile this view with the fact that observed individual VA differences were often quite close to or less than zero. Thus, the standard deviations observed for baseline VA gradients in both the major depressive disorder and control groups were similar to or greater than the means, and 6 of 11 patients with major depressive disorder appeared to have no VA gradient after selective serotonin reuptake inhibitor (SSRI) treatment. It seems inconceivable that a substantial fraction of subjects has little or no brain serotonin turnover and that others have many-fold higher rates of turnover. Even putting aside those subjects with zero or negative VA differences, there is a tremendous interindividual range that is inconsistent with the underlying physiological process that is purportedly indexed.

Barton and colleagues go on to suggest that the group mean reduction they observed in VA gradients for 5-HIAA after SSRI treatment is consistent with prior reports of reduced human lumbar cerebrospinal fluid (CSF) 5-HIAA levels after such treatment. However, the previous studies of human and nonhuman primate CSF 5-HIAA levels have been quite congruent in reporting mean reductions to between 40% and 60% of baseline, with small coefficients of variation.² In contrast, VA gradients observed by Barton and colleagues after SSRI administration ranged from zero to 140% of baseline (mean, approximately 33%), with a coefficient of variation of more than 100%. Parenthetically, when discussing the course of SSRI effects, Barton et al do not cite or take account of more recent studies suggesting that SSRI effects can occur quite quickly.^{3 - 4}

Finally, a prior report⁵ of elevated lumbar CSF serotonin levels in depression is twice cited by Barton et al¹ as relevant to discussions of serotonin functioning in depression. However, subsequent research⁶ convincingly demonstrated that actual levels of serotonin in control human lumbar CSF samples are usually less than 10 pg/mL, much lower than those reported by Gjerris and colleagues⁵ of about 65 pg/mL. As previously pointed out,⁶ concentrations of human lumbar CSF serotonin reported in the literature are uninterpretable because of problems with analytical specificity and unaccounted-for blood contamination and the extremely low endogenous levels. Serotonin levels of 50 to 100 pg/mL are present in cisternal CSF of nonhuman primates and appear to provide a useful index of serotonin release.²

In summary, because of apparent serious methodological limitations, the data presented in Barton et al¹ do not justify the conclusions expressed in the title and text regarding an elevation in brain serotonin turnover in depression. The data also should not be used to support inferences regarding the effects of genotype and therapy on central serotonin. Assessment of in vivo serotonin release and turnover remains a difficult task and a controversial area. Improved approaches are needed if the etiopathophysiological roles of serotonin in mood problems and related behaviors are to be clarified.

Correspondence: Dr Anderson, Yale Child Study Center, 230 S Frontage Rd, New Haven, CT 06520 (george.anderson@yale.edu).

Financial Disclosure: None reported.

Funding/Support: Cited research of Dr Anderson was supported by the National Institute of Mental Health and the Korczak Foundation for Autism Research.