(P-076) A Controlled 1-year Mirror-image Study of the Effect of Paliperidone Palmitate on Hospitalization in the Treatment of Schizophrenia in China

Background: The results are inconsistent between meta-analyses of randomized controlled trials and real-world mirror-image studies on the effectiveness superiority of long-term acting antipsychotic injections compared to oral antipsychotics.

Objectives: To evaluate whether paliperidone palmitate (PP1M) decreases hospitalization compared to oral atypical antipsychotics (OAA) in schizophrenia patients in China using a controlled 1-year mirror-image design to control for unmeasured confounders and time trend of hospitalization.

Methods: One meidcal database (2011-2018) of a psychiatric hospital in China was used. The PP1M group included adult schizophrenia patients who were prescribed PP1M in 2012-2017 and had used OAA in 1-year before PP1M, and had no other injectable antipsychotics and clozapine in 2011-2018. The OAA group included schizophrenia patients who were treated with OAA for at least 2 years and had switched OAA for at least once, but not treated with injectable antipsychotics and clozapine in 2011-2018; and were 1:1 randomly matched to PP1M patients on age, sex, and calendar year of OAA prescription. The index dates were defined as the date of the first PP1M and matched any OAA prescription, respectively. The outcomes include psychiatric hospitalization count and total bed days. The count was analyzed by conditional Poisson regression in either PP1M or OAA group to estimate rate ratio (RR) between post- and pre- index periods; the interaction between period (pre- vs. post-) and group (PP1M vs. OAA) was used to compare the difference on the change between different periods, a P-value for interaction term <0.05 indicates PP1M group difference between post- and pre- index period is statistically significant. Total bed days was tested by Wilcoxon matched-pair signed-rank test, a P-value for between group comparison <0.05 indicates a significant difference in PP1M group.

Results: 181 pairs of patients were selected, the average ages were 28.9±7.7 and 28.1±7.7 years in PP1M and OAA group, respectively. 47.5% were males in both groups. In PP1M group, 141 hospitalizations were in pre- and 13 in post- index period, RR=0.138 (95%CI=0.068-0.259); in OAA group, 33 vs. 48 in pre- and post- index period, RR=1.095 (95%CI=0.650-1.871); P for interaction was <0.0001. In PP1M group, average bed days were 33.3±36.9 in pre- and 2.9±13.4 in post- index period, P<0.0001; in OAA group, 5.1±20.9 vs. 7.2±43.2, P=0.515; P for between group comparison was <0.0001.

Conclusions: Subject to the limitation of the mirror-image design, PP1M was superior to OAA on the effectiveness of hospitalization count and bed days in Chinese schizophrenia patients.