We aimed to assess treatment persistence rates with LUR in a group of patients with schizophrenia treated in real-world Italian clinical practice and found that 78% were persistent to LUR treatment for at least 6 months and 58% were persistent for at least 18 months. Our persistence rate was relatively high compared to other studies such as the CATIE trial [1], where only about 25% (371 of 1432) of the patients who received at least one dose of study medication (oral olanzapine, perphenazine, quetiapine, risperidone, or ziprasidone) were persistent to their oral treatment for at least 18 months.

In an observational, prospective follow-up of 69 patients consecutively prescribed lurasidone in a large inner-city National Health Service mental health trust in London [3], 45 patients (65%) were known to have discontinued lurasidone at 1 year. Interestingly, most non-persistent patients discontinued LUR because of a perceived inefficacy, which might be related to the unusually high proportion of treatment-resistant subjects in the study sample [6].

In fact, while none of the clinical and demographic variables considered in our study had a statistically significant relationship with LUR treatment persistence, in the study conducted in London, patients who were not treatment-resistant had a substantially reduced risk of discontinuation. Of interest, higher LUR doses were a positive predictive factor for treatment persistence [6].

A study involving 146 Medicaid insured (52.1% male, mean age 43.5 years) and 63 commercially insured (46.1% male, mean age 40.0 years) patients treated with lurasidone evaluated variables such as medication possession ratio (MPR), mean time to discontinuation, and discontinuation rate [7]. In the Medicaid group, patients receiving lurasidone exhibited a significantly lower discontinuation rate compared to those treated with all other antipsychotics (49.3% versus 62.3–68.3%, all p < 0.05), the mean time to discontinuation with lurasidone was significantly longer than with ziprasidone (p < 0.05), and the MPR was 0.60 for lurasidone, versus 0.41–0.48 for other antipsychotics (all p < 0.05). In the commercially insured population, the discontinuation rate (44.4%) was lower for patients on lurasidone compared to patients on all other antipsychotics except risperidone (p < 0.05), the mean time to discontinuation was longer for lurasidone compared to other antipsychotics, and the MPR for lurasidone (0.61) was higher than the MPR for quetiapine (0.44) and ziprasidone (0.43) (both p < 0.05) [7].

The relatively high persistence rate to LUR in our and other studies may reflect LUR’s relatively favorable balance between efficacy and tolerability. Indeed, LUR has a relatively favorable metabolic side effect profile, and has been recommended as one of the choices when hoping to reduce the risk of metabolic adverse events [8, 9]. For instance, in a double-blind, 12 month, active-controlled study testing the long-term tolerability and safety of lurasidone in schizophrenia [10], a higher proportion of subjects treated with risperidone had an increase in weight of at least 7% compared to LUR (14% for risperidone vs. 7% for lurasidone). Also, the median increase in prolactin was significantly higher for risperidone (p < 0.001), while a similar improvement in efficacy measures and similar rates of relapse were observed with both medications. All-cause discontinuation rates were higher for lurasidone versus risperidone but long-term treatment with lurasidone was generally well tolerated.

Interestingly, an improvement in schizophrenia severity, according to the CGI-S, was observed in this study, as highlighted by the reduction in the proportion of patients being defined markedly ill and severely ill, as well as the increased percentage of patients being defined borderline mentally ill and mildly ill, following LUR therapy, compared to before LUR initiation.

In our study, lurasidone was relatively well tolerated, with at least one adverse drug reaction possibly related to LUR being reported for 6 patients (14.63%). ADRs consisted of akathisia (7.32%), extrapyramidal disorder (4.88%), hyperprolactinemia (2.44%), restlessness (2.44%), and galactorrhea (2.44%). None of the events were serious.

The main limitations of this study include its observational retrospective nature, which may bring the risk of patient selection bias, incomplete or missing data, lack of internal validity (no control group), difficulty in interpreting or verifying documented information, and variability among patients in the quality of documentation. Also, the information collected in the eCRF was limited to what was available in the medical records at the participating centers and did not include data related to health services received outside of that setting.

Additionally, the study was conducted in 5 Italian sites and the generalizability of our findings in other healthcare systems is difficult to ascertain. Moreover, the study sample size was relatively small, which may have prevented the observation of the impact on persistence by variables (e.g., rare side effects) that are infrequent. Finally, the lack of significant correlations between persistence and baseline demographic or clinical variables may have been simply due to the lack of statistical power, given the small sample size, and the fact that most patients were receiving concomitant medications may have had an impact on treatment persistence.

Nonetheless, a real life, non-interventional retrospective design allows the possibility to study conditions that are less likely to be influenced by the constraints of a more structured study design and that are often more representative of the patients and clinical scenarios that are encountered in the real-world practice.