**4. Discussion**

This study aims to analyze differences in circadian rhythmicity in patients under treatment with SUD attending to comorbid SMI, as well as its possible relationship with epidemiological and clinical characteristics.

Regarding sociodemographic and clinical results, the three groups of dual patients showed characteristics in line with previous studies [5,60–62]. Our results indicated an important presence of factors related to a worse clinical symptomatology and prognosis, especially for patients in the SZ+ group, and are consistent with the available literature [63–65]. Moreover, for patients with SZ+ and BD+, the outpatient treatment modality was predominant, while for patients with MDD+ the residential and therapeutic community treatment was more frequent. On the other hand, the mean age of SMI onset was earlier for the SZ+ group, this observation has been associated with a worse clinical, cognitive and functional prognosis [60,66,67]. As in previous studies, our sample presented similarities in psychiatric family history, medical disease comorbidity and previous suicide attempts [68–70]. Regarding nicotine consumption, although the three groups showed a high percentage of smokers and a moderate level of dependence, the SZ+ group exhibited the highest consumption and the MDD+ the lowest one.

Moreover, the longer duration of the SUD in patients with MDD+ may be related to the older age of the group, greater latency from the onset of depression until they seek professional help or may be unsuccessful attempts at previous treatments. According to available publications, in the three groups the most commonly used substances were cocaine, alcohol and cannabis [69,71,72], with an important common pattern of polydrug use in all the groups [4,60,69,72]. While the main substance of abuse may have a specific role in clinical and circadian rhythmic variables, polydrug use as a common pattern makes difficult to address such specific analysis. On the other hand, the severity of addiction was higher in the SZ+ group (followed by MDD+) in association with the consumption of a greater number of substances [73]. All these indicators would confirm the need for continuous, comprehensive treatments that affect relapse prevention [70].

Regarding the circadian typology, and in line with a previous study [15], in the MDD+ group the morning typology was the predominant one while in the SZ+ and BD+ groups the intermediate typology was the most frequent. The highest percentage of people with morning typology was observed in patients receiving residential treatment vs. outpatient modality. This observation is consistent with previous studies [14,15,43] and points out that there could be a possible regulating effect of the circadian rhythm generated by the habits and routines imposed by a residential treatment. The restoration of an adequate circadian rhythm is an element that contributes to the clinical improvement of patients with major depression [44], and according to our results, this also could be extended to those with MDD+, SZ+ and BD+.

On the other hand, in agreemen<sup>t</sup> with previous data in patients with a single SMI diagnosis, such as schizophrenia and bipolar disorder [74], the total duration of daily sleep in SZ+ patients were higher than those with MDD+, placing patients with BD+ in an intermediate position. This could be explained by the delay in getting up observed in patients with SZ+, which may be related to the sedative effect of most typical/atypical antipsychotic drugs and anticholinergics they were taking [66]. Likewise, those who received outpatient treatment slept more hours a day, went to bed and go<sup>t</sup> up later, and took more and longer naps. This suggests that sleep-wake rhythm time imbalances are not simply a consequence of clinical symptoms and could be influenced by treatment modality. The morningness tendency can be considered a marker of adherence to treatment and as a protective factor for relapses in both SUD and depressed patients [14,15,75]. Treatments that enhance synchronization with the environmental signals of the light-dark cycle, work on the regularity of schedules and include practice of physical exercise influence circadian recovery [19,25]. Our findings emphasize need to incorporate chronobiological adjustment strategies in dual patients under outpatient treatment modality, especially in the cases of those with SZ+ or BD+.

Regarding the circadian pattern of DST, it indicates less activation and/or greater daytime sleepiness in SZ+ patients and, to a lesser extent, also in BD+ patients. Even though there are no published data on DST in SZ+ patients, our findings are consistent with studies that have evaluated circadian functioning in patients with a diagnosis of schizophrenia only, where a lower amplitude and greater fragmentation of the activity-rest pattern were also observed [26,28,29]. In addition, we found a higher M5 value in BD+ patients compared to MDD+, that, together with a greater stability of the rhythm, points out a better night's rest [76] in BD+ patients. However, in the three groups of our sample this

data was found within normality values according to population norms. The inclusion of the HC group widened the differences found in the minimum and mesor values, Rayleigh, P12, IC, IV and M5. The SZ+ group obtained a significantly higher minimum and mesor value that denotes a lower diurnal activation [76] compared to control subjects and with the MDD+ group, without differences from BD+.

It is worth mentioning the differential relationship observed between tobacco consumption and circadian rhythmicity [14] regarding the comorbid SMI. For SZ+ patients nicotine dependence was associated to the quality of wakefulness, while for patients with MDD+ nicotine dependence was linked to the sleep period (M5). Thus, tobacco consumption and its level of dependence could be considered as a modulating factor of circadian rhythmicity, which is also related to the type of SMI diagnosis. Even though this should be deepened in the future, smoking seems to impair the quality of sleep for MDD+ patients. Furthermore, the poorer quality of wakefulness shown by patients with SZ+ is minimized in those who smoke, which could be explained by the palliative effect of nicotine over the side effects of antipsychotic treatment.

On the other hand, we found more stability of the circadian rhythm (Rayleigh vector and the power of the first 12 upper harmonics) in the MDD+ group than in the HC group. A previous study [15] observed that superior stability occurred in patients in therapeutic community (residential treatment) vs. those who were in an outpatient program. Therefore, if we take into account that the majority of patients in the MDD+ group underwent treatment in therapeutic community, our results could be congruen<sup>t</sup> due to the probable influence of the type of treatment on circadian rhythmicity. A result to emphasize is the lower rhythm stability index (IC) for patients with SZ+ compared to the HC group, which has been related to a more immature circadian system [58]. Furthermore, the three groups of patients showed a lower IV compared to the HC group. Furthermore, less fragmentation was also found for both the SUD and in the MDD+ groups, regardless of treatment modality [15]. Despite the absence of previous data about rhythm fragmentation in SZ+ and BD+ patients, the alteration of the IV in both cases suggests that it could be used as a psychopathological marker associated with SMIs, as it has been observed in schizophrenia [26,28] and bipolar disorder [23,39] conditions, regardless of the presence of a comorbid SUD.

Regarding treatment modality, our results sugges<sup>t</sup> a better quality of both sleep and wakefulness and a more robust circadian pattern of DST in dual patients under residential treatment. The patients under outpatient treatment, however, showed less daytime activation (minimum, higher L10 value and mesor) and a more evening pattern (later acrophase) compared to patients in residential facilities. Overall, these observations reveal a low contrast in their day-to-day life [77] in consistency with previous observations made in MDD+ [15] and in depression without SUD [44]. Therefore, it is emphasized that the treatment of dual patients, regardless of their SMI comorbidity, should promote rhythmic organization, physical activity in the open air and stable feeding times to maximize a good circadian expression and a morning pattern [14,55,78].

This work has some limitations, such as the cross-sectional design with a sample composed only by men, which rules out the establishment of causal relationships and does not allow the generalization of the results to women. Likewise, the wide age range of the sample, although partially controlled with age as a covariate, might have contributed to type II error. The high pattern of polydrug use in the patients does not allow us to assess specific associations between the type of substance and circadian rhythmicity. On the other hand, the higher proportion of patients in residential treatment in the MDD+ group could have influenced some of the circadian rhythm results attributed to their diagnosis. Future studies should evaluate circadian rhythmicity at the beginning as well as during treatment in order to know the differential evolution of patients and to identify possible risk factors and predictors of therapeutic adherence.
