We read with interest the article by Cox et al ^{Reference Cox, Benkelfat, Dagher, Delaney, Durand and Kolivakis1} and the insightful editorial by Nutt ^{Reference Nutt2} and applaud both the research staff and the patients involved in this important study in view of the ethical issues and challenges in their work. They provide supportive evidence that low serotonin activity can increase dopaminergic responses to cocaine in humans, suggesting a possible mechanism specific to ‘a low-serotonin state’ in causing addictive behaviours. Although illuminating, the results of the study should be interpreted with caution.

First, Cox et al use acute tryptophan depletion producing a reduction in plasma tryptophan, assumed to represent low levels of serotonin in the brain. The primary neuropharmacological effect of cocaine is to block the uptake of monoamines released into synapses, thereby increasing synaptic monoamine availability. It has been shown that cocaine can increase extracellular levels of serotonin in the nucleus accumbens of rats. ^{Reference Teneud, Baptista, Murzi, Hoebel and Hernandez3} Notably, in Cox et al's study, plasma concentrations of tryptophan did not significantly differ between cocaine and placebo, which appears to be an unexpected finding. This should be left open to discussion. Second, the interplay between serotonin and cocaine may be altered after repeated cocaine administration, ^{Reference Filip, Bubar and Cunningham4} a common manifestation in ‘real-world’ cocaine users. In this context, a study using an acute tryptophan depletion method plus repeated cocaine administration for patients with or without cocaine dependence, although ethically challenging, may obviously be of great clinical significance. Third, using repeated measures ANOVAs, it was assumed that the effects of cocaine did not carry over across conditions. Thus, it would have been clearer if the intervals between each condition were defined.

In addition to the issues raised by Nutt, ^{Reference Nutt2} as to the differences in response to various drugs of addiction, we would like to suggest that future research in the field of addiction focuses on using the tryptophan depletion test. For example, we now know that in pathological gamblers, dopamine release in ventral striatum correlates with excitement levels during the Iowa Gambling Task. ^{Reference Linnet, Moller, Peterson, Gjedde and Doudet5} However, tryptophan depletion significantly decreased, rather than increased, the number of decisions made to chase losses and the number of consecutive decisions to chase, independent of changes in mood. ^{Reference Campbell-Meiklejohn, Wakeley, Herbert, Cook, Scollo and Ray6} These findings, indubitably, did not support the hypothesis that low serotonin transmission may predispose to increased susceptibility to impulsive behaviours. It would be of interest to investigate the extent to which tryptophan depletion regulates dopamine release in patients who gamble, or in other populations with addictive behaviours, such as internet addiction or sex addiction, under control of the relevant stimuli.

Finally, in the editorial by Nutt, ^{Reference Nutt2} it is unclear as to the statement ‘It seems that this might be the case as in Cox et al's study lowering 5-HT function by tryptophan depletion led to a reduction in the actions of cocaine to release dopamine that was to some extent paralleled by a reduction in cocaine craving.’ The study by Cox et al showed that low serotonin activity augmented, rather than diminished, dopamine release in response to cocaine.

In summary, Cox et al's study ^{Reference Cox, Benkelfat, Dagher, Delaney, Durand and Kolivakis1} is a valuable contribution to the field of addiction, and we anticipate further studies examining the relationship between experimental reductions in serotonin activity and endogenous dopamine release in various addictive behaviours under control of the relevant stimuli.