Sir

We appreciate the comments of Dr Monteleone and colleagues and share their interest in the hypothesis that variations in leptin signaling early in the course of antipsychotic treatment could contribute to variance in weight gain observed during treatment. They mention their own previously published study that detected a significant inverse correlation between plasma leptin early in the course of treatment and subsequent weight change, with higher leptin levels associated with lower weight increases (Monteleone et al, 2002). However, neither their study nor the studies cited in their paper included the control group necessary to test the implicit hypothesis mentioned above.

By way of background, individuals are hypothesized to experience a counter-regulatory increase in plasma leptin during weight gain, based on the established role of leptin as a satiety signal in rodents (Schwartz et al, 2000). The missing studies that we mention in our paper, which have not been reported to date, would address the question of whether leptin signaling during antipsychotic-induced weight gain is appropriate for the level of adiposity achieved by treated individuals. It is likely that many people experience counter-regulatory plasma leptin increases during weight gain that serve to discourage further weight gain. To test the hypothesis that antipsychotic weight gain is secondary to abnormal leptin secretion or signaling, one could perform a study consisting of prospective randomized assignment to an antipsychotic drug treatment expected to produce weight gain, to a negative control condition (placebo or antipsychotic not associated with weight gain), and, most importantly, to a positive control condition such as excess feeding with no change in medication. The comparison of interest would be between the two conditions where weight increased, in order to evaluate whether leptin signaling (eg crudely measurable in vivo in humans via plasma leptin) is over- or underelevated as compared to plasma leptin elevations seen during weight gain due to simple increase in caloric intake. If plasma leptin was underelevated compared to excess feeding, this might suggest that antipsychotic treatment may interfere with leptin secretion or clearance, in either case interrupting normal leptin signaling that might otherwise reduce treatment-associated weight gain. With sufficient sample size, one could further evaluate whether individual differences in plasma leptin during antipsychotic-induced weight gain could help to explain which subjects experience more or less weight gain. Previous reports have not offered either the important control conditions or sufficient sample size to appropriately test these questions.

Unfortunately, all these suggested studies would be time consuming and expensive to conduct. Given our inability to detect inappropriate plasma leptin levels in antipsychotic-treated patients in comparison to adiposity-matched controls, the case for undertaking these subsequent studies is weakened (Haupt et al, 2005). As stated in our paper, along with the recent report by Herran et al, evidence does not currently support the hypothesis that leptin regulation in schizophrenia patients during antipsychotic treatment, as compared to leptin regulation in untreated healthy controls, is due to impaired leptin secretion or leptin resistance (Haupt et al, 2005; Herran et al, 2001).