To the Editor: We thank Drs. Meltzer and Shim for their thoughtful comments. Dr. Meltzer raises a valid issue that sertindole’s k_off is rather close to the range for typical antipsychotics, although the drug is an atypical antipsychotic. This is a valid concern; however, it is not just the k_off of the drug, but also its in vivo occupancy range, that is critical for atypicality. In that regard, sertindole at 8–24 mg/day occupies 61%–72% of D₂ receptors (1, 2), which should give rise to response without extrapyramidal symptoms. Further, a patient taking sertindole, 20 mg/day, who had also received 100 mg of fluphenazine decanoate 52 days before brain imaging had 86% D₂ occupancy. This patient exhibited akathisia (2), despite the fact that the patient would have had high 5-HT₂ occupancy.

We do not suggest that 5-HT₂ receptors make no contribution to the overall picture of atypicality. We propose that 5-HT₂ receptors are not necessary for atypicality, as shown by studies of amisulpride, remoxipride, and now aripiprazole, and that they are not sufficient for atypicality, as shown by studies of MDL-100907 and fananserin. Even in drugs with mixed 5-HT₂ and D₂ action, D₂ receptor affinity is the biggest factor for atypicality. In fact, Meltzer et al. (3) stated, "The low D₂ affinity values of atypical drugs contribute more than their high 5-HT₂ affinity values" (emphasis added); 64% of the typical-atypical difference, as revealed by discriminate function analysis, was accounted for by low affinity at the D₂ receptor, and only 17% was accounted for by the addition of 5-HT₂ receptor affinity (3).

Dr. Shim questions whether a slow k_on and low affinity could also give rise to atypicality. The affinity of a drug is the ratio k_off/k_on (not k_on/k_off, as Dr. Shim suggests) (4). While current differences among atypical antipsychotics are driven mainly by k_off, it is possible in theory that if k_off were held constant, there could exist a low-affinity drug by virtue of a very slow k_on. However, it is important to understand that k_on and k_off are rate constants (not the rates of association and dissociation). The rate of association is a function of k_on multiplied by the concentration, whereas the rate of dissociation depends on k_off alone. Since low-affinity drugs are given at much higher doses than high-affinity drugs, even if the k_on (the association rate constant) were low, the rate of association would still be high because of the high concentration. This is why we think that low affinity and fast k_off, and not low affinity and slow k_on, is important to atypicality.

Thus, "atypicality" should be seen as a continuum rather than as a dichotomy. Antipsychotics become increasingly more atypical (i.e., have less chance of extrapyramidal symptoms) as their affinity at the D₂ receptor decreases and their k_off increases. We are not suggesting that antipsychotics with no D₂ affinity are not possible, only that there are none at present.