The antioxidant defense system seems to be decreased in our population of children and adolescents with a first early onset psychotic episode and who were either treatment-naïve or had been treated with antipsychotics for only a short time when the blood samples were taken at admission. In particular, we found a decreased total antioxidant status and lower GSH levels in patients compared to healthy controls. Furthermore, compared with the control group, TAS was significantly lower in each of the three main diagnostic groups (psychotic disorder not otherwise specified, bipolar disorder and schizophrenia), but GSH was significantly lower only in the schizophrenia group. Moreover, for the oxidative stress variables measured, a between-diagnostic group difference was significant only for TAS, with the lowest TAS levels in the schizophrenic and bipolar patients. From the baseline data it appears that patients with more severe and chronic diseases (SCH and BIP) have more oxidative stress than those with acute psychosis. Indeed, it must be taken into account that although the oxidative data shown were obtained at baseline, the diagnoses were only confirmed during the follow-up. These preliminary findings suggest that antioxidant problems may be a function of diagnosis, in this case schizophrenia and bipolar disorder.
Lipid oxidative damage seems to be increased in early first psychotic episode patients, measured as elevated levels of lipid hydroperoxides (LOOHs); the activity of cGPx is also elevated in patients compared with healthy controls. When comparing the patient subgroups by confirmed diagnoses with the control group, only bipolar patients had elevated LOOHs.
There is no bias in the results due to treatment use, because there were no differences in oxidative stress when we compared the patients who were treatment-naïve with those who had received antipsychotic treatment before admission. Previously, it has been stated that psychopharmacological treatment can partially reverse oxidative stress . However, in our patient sample, treatment was used for only a short period before hospitalization, and we have not yet performed a longitudinal prospective analysis of the sample. We found a relationship between age and oxidative stress when the patients and controls were considered together. However, we cannot conclude that the older the age, the higher oxidative stress, as our older patients were adolescents. Other studies have not found any age-related differences in total antioxidant status in healthy volunteers .
A defect in antioxidant protection seems to be a common feature of patients with a first psychotic episode, but has been little explored so far in early onset psychosis and patients who have been psychotic for only a short time. This defect should be taken into account when developing neuroprotective treatments.
Despite the conflicting data found previously [40–43], abnormalities in antioxidant enzyme activities are frequently found in patients diagnosed with psychosis or bipolar disorder. This imbalance in enzyme activities reduces the efficacy of the cellular antioxidant defense, which requires a proper balance of antioxidant enzymes, and contributes to brain pathology . Indeed, we found a significant correlation between SOD and LOOH levels indicating that oxidative stress is probably producing lipid damage. It has been taken into account that the brain contains almost no catalase and less cGPx than other tissues. Although these enzymes may serve as peripheral indicators of oxidative stress, this makes our results for GSH levels and SOD activity more interesting.
The GSH deficit found in this study and in previous reports [15, 40, 45–47] may be involved in membrane peroxidation and microlesions related to dopamine, which seem to be increased in psychosis, and suggest that GSH may be a possible indirect indicator of damage in neuronal membranes [48–52]. Anomalies in GSH metabolism were also supported by the low expression of the gene of the key GSH-synthesizing enzyme, glutamate cysteine ligase modifier subunit, in patient fibroblasts . Moreover, inhibition of brain glutathione synthesis and dopamine uptake in developing rats induces long-term cognitive deficits in adulthood . Therefore, it seems that dopamine, which has a neurotoxic potential, contributes to cellular oxidative stress, which can be exacerbated if glutathione synthesis is compromised . In addition, the concentrations of glutathione and one of its metabolites, glutamylglutamine, are reduced in the cerebrospinal fluid of drug-free patients . An animal model of schizophrenia has recently been proposed where a redox imbalance during postnatal development induces abnormalities in cortical development . Also, a recent study performed in cultured fibroblasts of patients with schizophrenia found impaired genetic and functional capacity to synthesize GSH under conditions of oxidative stress in schizophrenic patients . These converging data, in agreement with our results in child and adolescent patients, indicate that psychosis is associated with an important brain glutathione deficit. One question not answered by this study is why early-onset bipolar patients (who also have dopamine metabolism alterations) do not show a significant GSH deficit. It is possible that patients diagnosed with bipolar disorder share some etiological and physiopathological mechanisms with schizophrenia, but not all. Indeed, patients with bipolar disorder usually have less cognitive dysfunction in the long term and better prognosis than patients with schizophrenia. In fact, it could be hypothesized that different etiological mechanisms converge into precipitating a psychotic episode in an adolescent, and only in individuals with a limited GSH synthesis capacity, after which the psychotic episode develops into a degenerating condition that we call schizophrenia. This could be tested by analyzing glutathione in high risk populations that are subsequently followed up.
In our patient group, the presence of manic symptoms was associated with lower antioxidant capacity. Previously, Andreazza et al.  demonstrated that DNA damage, probably due to oxidative stress, was increased in mania patients and was related to severity of mania. We also found that antioxidant defense was related to global functioning.
Oxidative stress has been related to DNA damage, and it has been suggested that psychopharmacological treatment can partially reverse oxidative stress. As oxidative stress could be a marker of severity in patients with medical conditions such as atherosclerosis  or diabetes [61, 62], long-term investigations should determine whether it is also a marker of severity in patients with psychosis. It is possible that oxidative stress could also mediate the vascular damage of patients with psychosis as it has been established that patients with psychosis are at higher risk of developing cardiovascular diseases  and diabetes .
Although we should be cautious, our findings support the possibility of using peripheral markers of oxidative and antioxidative balance in patients with first-episode psychosis, taking into account the special sensitivity of the brain to oxidative damage .
This study has some limitations. Data are cross-sectional, diagnostic subgroups are relatively small, and it was difficult to establish in advance a sample size to perform the data analyses because of the paucity of studies with similar design characteristics; nevertheless, our patient sample is one of the largest reported to date. A second limitation is the type of recruitment centers; the majority was hospital settings with an inpatient facility and this may represent a bias towards the inclusion of more severe cases, making it difficult to generalize the results to less severe ones. Because the data presents changes in peripheral blood, further work is needed to determine if such changes adequately reflect changes in the brain and of mental state. The strengths of the study are the uniformity in age with an early onset and first episode of all psychoses, and the existence of a control group.