Fluoxetine, paroxetine, and many other SSRIs have proven to be efficacious for the treatment of adults with depression, premenstrual syndrome, obsessive compulsive behavior, panic disorder, generalized anxiety disorder, social anxiety disorder, and post-traumatic stress disorder [1–3]. These drugs have been found to be safer and have fewer side-effects than tricyclic antidepressants. SSRIs have also been used in children and adolescents, and fluoxetine is approved for use in those 6-18 years of age. In children whose nervous system is still developing, there are concerns about the long-term effects of these drugs.
In a recent preclinical experiment examining paroxetine and fluvoxamine, it was reported that these drugs reduced time-in-open in an EPM . Decreased time in open in the EPM is the most widely accepted index for this test of increased anxiety-like behavior . In the experiment by deJong et al.  this change occurred for both drugs 20-30 days after the end of treatment. No change was seen in the number of zone crossings. These authors also tested the animals in an elevated T-maze (the EPM was used with one arm blocked). Out of 5 trials for time to enter an open arm after being placed in a closed arm, both SSRI-treated groups entered an open arm faster on trial-2 compared with controls but not on trial-1 or 3-5. When the situation was reversed and they were given an additional trial by being placed in an open arm and timed for entry into a closed arm, no differences were found. No differences in ASR/PPI were obtained nor were changes obtained on immobility time in the FST.
We sought to further test the effects of adolescent exposure to SSRIs and followed many of the design features used previously , except using more dose levels of each drug and including separate groups of animals, half tested while on-drug (Group A) and half tested off-drug (Group B; 60 days post-treatment). If long-term effects of SSRI treatment were obtained, it would potentially represent a concern for the safety of the drugs.
Both drugs produced body weight reductions at the highest doses tested by P44, i.e., after 12 days of treatment, and these reductions remained through the end of treatment (P62). But the effects showed recovery to non-significant differences one week post-treatment. Behaviorally, no treatment-related effects on the EPM were obtained on the principal measure of anxiety, i.e., time-in-open. In addition, no differences were found for number of zone crossings or latency to first open arm entry. The Group A Par17 animals showed a significant reduction in head-dips, an effect seen in no other group, however this effect was not seen in Group B when tested 60 days after drug cessation. deJong et al.  did not report head-dips, therefore, it is not possible to make a direct comparison on this variable, however, a reduction in head-dips is generally interpreted as an increase in anxiety , although it is an index of anxiety that is secondary to time-in-open and therefore provides less persuasive evidence of a significant change in anxiety. However, this effect would be consistent with the finding of deJong et al. . A recent study showed that rats treated from P25-49 with 12 mg/kg of fluoxetine had reduced time-in-open in the EPM  7 days post-treatment, a finding consistent with reduced head-dips that we found while animals were on drug. It is unclear why adolescent exposed rats and mice show greater inconsistency in EPM responses to SSRIs than adult rodents. Some of it may be related to the different exposure ages as there is no consistent definition of adolescence in rodents. Doses and routes of drug administration as well as duration of treatment and end of treat to test interval vary widely across studies, increasing the difficulty of discerning patterns.
Neither we nor deJong et al. (2006) found any effect on the FST test of swimming despair [17, 18] and this finding is different than that reported by Homberg et al.  but they treated rats earlier (P25-46); moreover, they found no differences in rats treated at a later age. Other studies have found opposite or paradoxical effects of adolescent exposure to fluoxetine in mice, but only in one of the two strains tested . To the extent that the FST is a valid preclinical test of depression, the data do not suggest that adolescent exposure to fluoxetine or paroxetine result in long-term changes in immobility. This is consistent with a newer report of chronic fluoxetine treatment in mice from P14-42 showing no FST effects  and the data cited above . We scored the FST for immobility time as originally described by Porsolt et al. [17, 18]. We did not score the test using the recently suggested indices of active swimming and wall climbing . While we cannot rule-out the possibility that these additional indices might have uncovered other effects, we did not include them because our interest was not differentiating among different classes of SSRIs, which was the reason these measures were introduced, but rather whether immobility was a long-term consequence of drug exposure after adolescent treatment.
Others have administered SSRIs by osmotic pump in order to maintain plasma concentrations in animals within the same range as human therapeutic concentrations [27, 28]. The intent of these studies was different than ours. These studies were designed to assess the molecular targets of SSRIs. For this purpose maintaining a constant drug concentration is desirable but this goal was not necessary for our purposes of determining if these drugs have persistent long-term effects long after cessation.
deJong et al.  reported no changes in ASR/PPI. We replicated their parameters and in agreement, observed no changes in ASR or PPI from either drug 60 days post-treatment. However, while on-drug, we saw a trend toward ASR facilitation that did not interact with PPI. To further explore the data, we conducted follow-up analyses by combining the two fluoxetine groups into one pooled group by averaging the data for the fluoxetine animals in each litter together to create a single, merged group. We did the same for the paroxetine, merging the data from all three dose levels together among littermates to create a single paroxetine group. Two follow-up analyses were performed. These were: (1) with all prepulse trials included and (2) with no-prepulse trials included, i.e., only the unmodified, basic ASR trials. Both analyses resulted in the same finding: the pooled fluoxetine and pooled paroxetine groups exhibited increased ASR amplitude compared with Controls. By contrast, when Group B ASR-PPI data were analyzed the same way, no residual effect or trend toward an effect was observed. This is in contrast to a study that found that P24-46 fluoxetine reduced ASR at a higher dose than we used (12 mg/kg), but similar to our data, found no effects at later ages .
The present data should be viewed within the limitations of the study. As in the deJong et al.  and Norcross et al.  studies, we tested only males. It may be worthwhile to test females as many drugs exhibit sexually dimorphic responses. It is also worth considering that future experiments might test a broader range of doses in order to more thoroughly test for possible adverse effects. Finally, other behavioral tests might be worth considering; Norcross et al.  also included open-field and cued fear conditioning with extinction, but even more tests might be considered such as fear-induced acoustic startle facilitation. There are many tests of anxiety, conditioned fear, sensorimotor gating, and depression, of which only a subset were used here or in the other studies cited that tested for residual effects. A more extensive battery of tests might reveal effects not detected herein.
Overall, the data show that for the doses of fluoxetine and paroxetine tested by an oral route during adolescent to early adult brain development (operationally defined as P33-62), caused minor and only transient reductions in body weight gain, a small but significant ASR facilitation during treatment that did not remain 60 days post-treatment, and a high-dose only (Par17) reduction in EPM head-dips while on-drug but not off-drug. The findings, when considered in terms of the power to detect differences in the present experimental design (≥ 20 animals per group), multiple dose levels, use of a within-litter design that improves subject matching, and inclusion of both on- and off-drug cohorts, suggest that there is no signal of adverse effects present in the data that might raise concern over the long-term safety of the drugs when treatment is during an interval spanning adolescent brain development in rats, in agreement with previous findings in mice . It was recently reported that fluoxetine given by continuous infusion via minipumps from P14-42 results in anxiogenic effects in mice while on-drug on tests of novelty-induced hypophagia in Swiss-Webster (SW) and C57BL/6 Charles River (B6) mice, however there were no effects in the EPM in the SW strain and reduced time-in-open in the B6 strain but no effects on open-field center time or FST in either strain. These effects disappeared off-drug  which is entirely consistent with the present findings. Thus, while SSRIs induce relatively reliable effects on-drug in adult rodents, their effects from adolescent exposure are less consistent. Our data are consistent in that we saw only small effects while on drug and no long-term effects long after drug discontinuation. The latter provides some evidence that these drugs are not neurotoxic when given during adolescent stages of brain development at least on the behavioral indices used here.