We found that the psychiatric population treated with antipsychotic monotherapy had much less risk of developing an increase in QTc interval compared to those treated with antipsychotics plus an antidepressant or lithium.
Two main mechanisms seem to operate in determining the prolongation of QTc interval during treatment with different combinations of psychoactive drugs. The first is the synergic blockade of the HERG potassium channels, the second is the increase in drug levels (with subsequent augmented risk of cardiotoxicity) due to metabolic interactions between drugs that share the same metabolic pathway . This mechanism may be particularly relevant in subject with genetic-determined impairment of CYP2D6 and CYP3A4 drug-metabolizing enzymes (poor metabolizer subjects) .
In our study, metabolic interactions leading to abnormal elevation of serum levels of antipsychotics did not seem to be the principal determinant of the greater QTc prolongation in the group with combined therapy. Indeed serum levels of antipsychotics were all within or under the expected range after therapeutic dosing in both groups, they were not higher in Group 2 compared to Group 1 and the highest prolongations observed were not associated with the highest antipsychotic serum levels.
Recently, Harringan et al  analyzed, in a prospective randomized study, the effects of six antipsychotics on the QTc interval; they found that each of the antipsychotics were associated with measurable QTc prolongation which was not augmented by concomitant use of metabolic inhibitors, even if in their study plasmatic levels of antipsychotics raised after the addition of the specific metabolic inhibitor.
In our study, the combination of different drugs doesn't seem to cause strong interactions on drug metabolism. However, in our sample, the combination of drugs that specifically interfere in their own methabolism, like fluvoxamine and olanzapine, paroxetine and risperidone, were avoided by clinicians.
Antidepressant used in our study have a mild inhibitory action on antipsychotic methabolism and this can explain why antipsychotic serum levels didn't raise in Group 2 compared to Group 1. It is reassuring to find that significant pharmacokinetic interactions do not occur when the antipsychotics studied were coadministered with antidepressant commonly used in clinical practice.
If the metabolic interactions do not seem to be the most important explanation for our results, an alternative explanation might be the synergic actions of different drugs on ion channels.
The two patients with the highest prolongations were both taking risperidone, which is noted to block the Ikr current [23, 24]. Actually, many psychotropic drugs share this capacity to inhibit Ikrcurrent, including not only antipsychotic agents but also antidepressant agents like citalopram, fluoxetine, paroxetine [16, 23]. Those agents may have synergic effect when used in combination.
This study has several limitations, most of them related to the naturalistic setting of this study: we chosed to administer the second ECG after four days of therapy at full dosage (generally after one week from recruitment) because all antipsychotic used reached the steady-state in 3–5 days. Actually we couldn't chose a longer interval between the first and the second ECG because the average duration of recovery in our ward is 8,5 days. Dosing was clinically determined for symptom response by the treating psychiatrist and hence, doses varied within and between groups. Moreover statistical comparison of QTc interval changes among agents was not possible because of the small number of the samples.
Finally we didn't measure antidepressant serum levels. Consistently with data reported in literature, we thought that antipsychotic would have been the principal drugs involved in QTc prolongation, while antidepressant would have only a role of potentiating agents. Actually, serum levels of antidepressants would have helped to explain the greater prolongation observed in Group 2.